feat: unified org-garden

feat: unified watch server under org-garden
feat: pipeline - org2md file watch
2026-02-21 13:10:30 +01:00 · 2026-02-21 00:36:31 +01:00 · 2026-02-20 23:31:06 +01:00 · 2026-02-20 17:54:12 +01:00 · 2026-02-20 15:08:04 +01:00 · 2026-02-20 11:11:42 +01:00
91 changed files with 3157 additions and 7312 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -9,3 +9,19 @@ tsconfig.tsbuildinfo
 private/
 .replit
 replit.nix
+erl_crash.dump
+# content/ is generated by the export script; only keep the placeholder
+content/*
+!content/.gitkeep
+# static/ox-hugo/ is populated by ox-hugo during export
+static/ox-hugo/
+# Elixir/Mix build artifacts for the pipeline project
+scripts/pipeline/_build/
+scripts/pipeline/deps/
+scripts/pipeline/erl_crash.dump
+# Test helpers (not needed in production)
+scripts/test.bib
+scripts/test_pipeline.exs
+/org-garden/deps/
+/org-garden/_build/
+/org-garden/result
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -0,0 +1,254 @@
+# AGENTS.md - Coding Agent Instructions
+
+This document provides essential information for AI coding agents working in this repository.
+
+## Project Overview
+
+**Quartz** is a static site generator for publishing digital gardens and notes as websites.
+Built with TypeScript, Preact, and unified/remark/rehype for markdown processing.
+
+| Stack         | Technology                                |
+| ------------- | ----------------------------------------- |
+| Language      | TypeScript 5.x (strict mode)              |
+| Runtime       | Node.js >=22 (v22.16.0 pinned)            |
+| Package Mgr   | npm >=10.9.2                              |
+| Module System | ES Modules (`"type": "module"`)           |
+| UI Framework  | Preact 10.x (JSX with `react-jsx` pragma) |
+| Build Tool    | esbuild                                   |
+| Styling       | SCSS via esbuild-sass-plugin              |
+
+## Environment
+
+This is a Nix project. Use the provided `flake.nix` to enter a dev shell with Node.js 22 and npm:
+
+```bash
+nix develop
+```
+
+All `npm` commands below must be run inside the dev shell.
+
+## Build, Lint, and Test Commands
+
+```bash
+# Type check and format check (CI validation)
+npm run check
+
+# Auto-format code with Prettier
+npm run format
+
+# Run all tests
+npm run test
+
+# Run a single test file
+npx tsx --test quartz/util/path.test.ts
+
+# Run tests matching a pattern (use --test-name-pattern)
+npx tsx --test --test-name-pattern="typeguards" quartz/util/path.test.ts
+
+# Build the static site
+npx quartz build
+
+# Build and serve with hot reload
+npx quartz build --serve
+
+# Profile build performance
+npm run profile
+```
+
+### Test Files Location
+
+Tests use Node.js native test runner via `tsx`. Test files follow the `*.test.ts` pattern:
+
+- `quartz/util/path.test.ts`
+- `quartz/util/fileTrie.test.ts`
+- `quartz/components/scripts/search.test.ts`
+
+## Code Style Guidelines
+
+### Prettier Configuration (`.prettierrc`)
+
+```json
+{
+  "printWidth": 100,
+  "tabWidth": 2,
+  "semi": false,
+  "trailingComma": "all",
+  "quoteProps": "as-needed"
+}
+```
+
+**No ESLint** - only Prettier for formatting. Run `npm run format` before committing.
+
+### TypeScript Configuration
+
+- **Strict mode enabled** (`strict: true`)
+- `noUnusedLocals: true` - no unused variables
+- `noUnusedParameters: true` - no unused function parameters
+- JSX configured for Preact (`jsxImportSource: "preact"`)
+
+### Import Conventions
+
+```typescript
+// 1. External packages first
+import { PluggableList } from "unified"
+import { visit } from "unist-util-visit"
+
+// 2. Internal utilities/types (relative paths)
+import { QuartzTransformerPlugin } from "../types"
+import { FilePath, slugifyFilePath } from "../../util/path"
+import { i18n } from "../../i18n"
+```
+
+### Naming Conventions
+
+| Element          | Convention   | Example                             |
+| ---------------- | ------------ | ----------------------------------- |
+| Files (utils)    | camelCase    | `path.ts`, `fileTrie.ts`            |
+| Files (comps)    | PascalCase   | `TableOfContents.tsx`, `Search.tsx` |
+| Types/Interfaces | PascalCase   | `QuartzComponent`, `FullSlug`       |
+| Type Guards      | `is*` prefix | `isFilePath()`, `isFullSlug()`      |
+| Constants        | UPPER_CASE   | `QUARTZ`, `UPSTREAM_NAME`           |
+| Options types    | `Options`    | `interface Options { ... }`         |
+
+### Branded Types Pattern
+
+This codebase uses branded types for type-safe path handling:
+
+```typescript
+type SlugLike<T> = string & { __brand: T }
+export type FilePath = SlugLike<"filepath">
+export type FullSlug = SlugLike<"full">
+export type SimpleSlug = SlugLike<"simple">
+
+// Always validate with type guards before using
+export function isFilePath(s: string): s is FilePath { ... }
+```
+
+### Component Pattern (Preact)
+
+Components use a factory function pattern with attached static properties:
+
+```typescript
+export default ((userOpts?: Partial<Options>) => {
+  const opts: Options = { ...defaultOptions, ...userOpts }
+
+  const ComponentName: QuartzComponent = ({ cfg, displayClass }: QuartzComponentProps) => {
+    return <div class={classNames(displayClass, "component-name")}>...</div>
+  }
+
+  ComponentName.css = style           // SCSS styles
+  ComponentName.afterDOMLoaded = script  // Client-side JS
+  return ComponentName
+}) satisfies QuartzComponentConstructor
+```
+
+### Plugin Pattern
+
+Three plugin types: transformers, filters, and emitters.
+
+```typescript
+export const PluginName: QuartzTransformerPlugin<Partial<Options>> = (userOpts) => {
+  const opts = { ...defaultOptions, ...userOpts }
+  return {
+    name: "PluginName",
+    markdownPlugins(ctx) { return [...] },
+    htmlPlugins(ctx) { return [...] },
+    externalResources(ctx) { return { js: [], css: [] } },
+  }
+}
+```
+
+### Testing Pattern
+
+Use Node.js native test runner with `assert`:
+
+```typescript
+import test, { describe, beforeEach } from "node:test"
+import assert from "node:assert"
+
+describe("FeatureName", () => {
+  test("should do something", () => {
+    assert.strictEqual(actual, expected)
+    assert.deepStrictEqual(actualObj, expectedObj)
+    assert(condition) // truthy assertion
+    assert(!condition) // falsy assertion
+  })
+})
+```
+
+### Error Handling
+
+- Use `try/catch` for critical operations (file I/O, parsing)
+- Custom `trace` utility for error reporting with stack traces
+- `process.exit(1)` for fatal errors
+- `console.warn()` for non-fatal issues
+
+### Async Patterns
+
+- Prefer `async/await` over raw promises
+- Use async generators (`async *emit()`) for streaming file output
+- Use `async-mutex` for concurrent build protection
+
+## Project Structure
+
+```
+quartz/
+├── bootstrap-cli.mjs    # CLI entry point
+├── build.ts             # Build orchestration
+├── cfg.ts               # Configuration types
+├── components/          # Preact UI components
+│   ├── *.tsx            # Components
+│   ├── scripts/         # Client-side scripts (*.inline.ts)
+│   └── styles/          # Component SCSS
+├── plugins/
+│   ├── transformers/    # Markdown AST transformers
+│   ├── filters/         # Content filters
+│   ├── emitters/        # Output generators
+│   └── types.ts         # Plugin type definitions
+├── processors/          # Build pipeline (parse/filter/emit)
+├── util/                # Utility functions
+└── i18n/                # Internationalization (30+ locales)
+```
+
+## Branch Workflow
+
+This is a fork of [jackyzha0/quartz](https://github.com/jackyzha0/quartz) with org-roam customizations.
+
+| Branch      | Purpose                                          |
+| ----------- | ------------------------------------------------ |
+| `main`      | Clean mirror of upstream quartz — no custom code |
+| `org-roam`  | Default branch — all customizations live here    |
+| `feature/*` | Short-lived branches off `org-roam`              |
+
+### Pulling Upstream Updates
+
+```bash
+git checkout main
+git fetch upstream
+git merge upstream/main
+git checkout org-roam
+git merge main
+# Resolve conflicts if any, then commit
+```
+
+### Working on Features
+
+```bash
+git checkout org-roam
+git checkout -b feature/my-feature
+# ... work ...
+git checkout org-roam
+git merge feature/my-feature
+git branch -d feature/my-feature
+```
+
+**Merge direction:** `upstream → main → org-roam → feature/*`
+
+## Important Notes
+
+- **Client-side scripts**: Use `.inline.ts` suffix, bundled via esbuild
+- **Isomorphic code**: `quartz/util/path.ts` must not use Node.js APIs
+- **Incremental builds**: Plugins can implement `partialEmit` for efficiency
+- **Markdown flavors**: Supports Obsidian (`ofm.ts`) and Roam (`roam.ts`) syntax
+- **Pipeline build artifacts**: `scripts/pipeline/_build/` and `scripts/pipeline/deps/`
+  are gitignored — run `mix deps.get` inside `scripts/pipeline/` after a fresh clone
--- a/README.md
+++ b/README.md
@@ -1,13 +1,96 @@
-# Quartz v4
+# Quartz v4 — org-roam edition

-> “[One] who works with the door open gets all kinds of interruptions, but [they] also occasionally gets clues as to what the world is and what might be important.” — Richard Hamming
+> "[One] who works with the door open gets all kinds of interruptions, but [they] also occasionally gets clues as to what the world is and what might be important." — Richard Hamming

 Quartz is a set of tools that helps you publish your [digital garden](https://jzhao.xyz/posts/networked-thought) and notes as a website for free.

-🔗 Read the documentation and get started: https://quartz.jzhao.xyz/
+This fork adds first-class support for [org-roam](https://www.orgroam.com/) notes via [ox-hugo](https://ox-hugo.scripter.co/).
+
+🔗 Upstream documentation: https://quartz.jzhao.xyz/

 [Join the Discord Community](https://discord.gg/cRFFHYye7t)

+## Quick Start
+
+### Prerequisites
+
+This project uses Nix. Enter the development shell, which provides Node.js 22, Elixir, and Emacs with ox-hugo:
+
+```bash
+nix develop
+```
+
+All commands below must be run inside this shell.
+
+```bash
+npm install
+```
+
+### Building from org-roam notes
+
+Your org-roam notes live in a separate directory. Point `NOTES_DIR` at it:
+
+```bash
+# Export notes to content/ and build the site
+NOTES_DIR=/path/to/notes npm run build:notes
+
+# Export, build, and serve with hot reload
+NOTES_DIR=/path/to/notes npm run serve:notes
+
+# Export only (wipes content/ and re-exports all .org files)
+NOTES_DIR=/path/to/notes npm run export
+```
+
+The export pipeline runs in four phases:
+
+1. **Wipe** `content/` clean
+2. **Export** every `.org` file via `emacs --batch` + ox-hugo → Markdown
+3. **Transform** — post-process the Markdown (citation resolution, etc.)
+4. **Index** — generate a fallback `index.md` if none was exported
+
+#### Citations (org-citar → Zotero links)
+
+org-citar references (`[cite:@key]`) are resolved to clickable Zotero links.
+With Zotero running and the [Better BibTeX](https://retorque.re/zotero-better-bibtex/)
+plugin installed, no extra configuration is needed — the pipeline detects it
+automatically and links directly to the PDF in your library.
+
+```bash
+# Use a local .bib file as fallback when Zotero is not running
+BIBTEX_FILE=/path/to/refs.bib NOTES_DIR=/path/to/notes npm run export
+
+# Control warning verbosity for unresolved keys
+CITATION_MODE=strict NOTES_DIR=/path/to/notes npm run export
+```
+
+| Env var         | Default                  | Purpose                                   |
+| --------------- | ------------------------ | ----------------------------------------- |
+| `BIBTEX_FILE`   | —                        | Path to `.bib` file for citation fallback |
+| `ZOTERO_URL`    | `http://localhost:23119` | Zotero Better BibTeX base URL             |
+| `CITATION_MODE` | `warn`                   | `silent` / `warn` / `strict`              |
+
+### Building without org-roam notes
+
+If you manage `content/` directly with Markdown files:
+
+```bash
+# Build the site
+npx quartz build
+
+# Build and serve with hot reload
+npx quartz build --serve
+```
+
+The site is generated in `public/`. When serving, visit http://localhost:8080.
+
+### Development
+
+```bash
+npm run check    # type check + format check
+npm run format   # auto-format with Prettier
+npm run test     # run tests
+```
+
 ## Sponsors

 <p align="center">
--- a/docs/features/Obsidian
+++ b/docs/features/Obsidian
@@ -10,10 +10,8 @@ By default, Quartz ships with the [[ObsidianFlavoredMarkdown]] plugin, which is

 It also ships with support for [frontmatter parsing](https://help.obsidian.md/Editing+and+formatting/Properties) with the same fields that Obsidian uses through the [[Frontmatter]] transformer plugin.

-Quartz also provides [[CrawlLinks]] plugin, which allows you to customize Quartz's link resolution behaviour to match Obsidian.
-
-For dynamic database-like views, Quartz supports [[bases|Obsidian Bases]] through the [[ObsidianBases]] transformer and [[BasePage]] emitter plugins.
+Finally, Quartz also provides [[CrawlLinks]] plugin, which allows you to customize Quartz's link resolution behaviour to match Obsidian.

 ## Configuration

-This functionality is provided by the [[ObsidianFlavoredMarkdown]], [[ObsidianBases]], [[Frontmatter]] and [[CrawlLinks]] plugins. See the plugin pages for customization options.
+This functionality is provided by the [[ObsidianFlavoredMarkdown]], [[Frontmatter]] and [[CrawlLinks]] plugins. See the plugin pages for customization options.
--- a/docs/features/bases.md
+++ b/docs/features/bases.md
@@ -1,42 +0,0 @@
---
-title: Bases
-tags:
-  - feature/transformer
-  - feature/emitter
---
-
-Quartz supports [Obsidian Bases](https://help.obsidian.md/bases), which allow you to create dynamic, database-like views of your notes. See the [official Obsidian documentation](https://help.obsidian.md/bases/syntax) for the full syntax reference.
-
-## Quick Example
-
-Create a `.base` file in your content folder:
-
-```yaml
-filters:
-  and:
-    - file.hasTag("task")
-
-views:
-  - type: table
-    name: "Task List"
-    order:
-      - file.name
-      - status
-      - due_date
-```
-
-Each view gets its own page at `<base-name>/<view-name>`.
-
-## Wikilinks
-
-Link to base views using the standard [[navigation.base#Plugins|wikilink]] syntax:
-
-```markdown
-[[my-base.base#Task List]]
-```
-
-This resolves to `my-base/Task-List`.
-
-## Configuration
-
-This functionality is provided by the [[ObsidianBases]] transformer plugin (which parses `.base` files) and the [[BasePage]] emitter plugin (which generates the pages).
--- a/docs/navigation.base
+++ b/docs/navigation.base
@@ -1,93 +0,0 @@
-filters:
-  and:
-    - file.ext == "md"
-formulas:
-  doc_type: |
-    if(file.hasTag("plugin/transformer"), "transformer",
-    if(file.hasTag("plugin/emitter"), "emitter",
-    if(file.hasTag("plugin/filter"), "filter",
-    if(file.hasTag("component"), "component",
-    if(file.inFolder("features"), "feature",
-    if(file.inFolder("advanced"), "advanced",
-    if(file.inFolder("plugins"), "plugin", "guide")))))))
-  last_modified: file.mtime.relative()
-  section: |
-    if(file.inFolder("plugins"), "plugins",
-    if(file.inFolder("features"), "features",
-    if(file.inFolder("advanced"), "advanced",
-    if(file.inFolder("tags"), "tags", "core"))))
-properties:
-  title:
-    displayName: Title
-  formula.doc_type:
-    displayName: Type
-  formula.last_modified:
-    displayName: Updated
-  formula.section:
-    displayName: Section
-views:
-  - type: table
-    name: All Documentation
-    groupBy:
-      property: formula.section
-      direction: ASC
-    order:
-      - file.name
-      - title
-      - formula.doc_type
-      - formula.section
-      - formula.last_modified
-    sort:
-      - property: formula.doc_type
-        direction: ASC
-      - property: file.name
-        direction: ASC
-    columnSize:
-      file.name: 185
-      note.title: 268
-      formula.doc_type: 146
-      formula.section: 276
-  - type: table
-    name: Plugins
-    filters:
-      or:
-        - file.hasTag("plugin/transformer")
-        - file.hasTag("plugin/emitter")
-        - file.hasTag("plugin/filter")
-    groupBy:
-      property: formula.doc_type
-      direction: ASC
-    order:
-      - file.name
-      - title
-      - formula.doc_type
-      - formula.last_modified
-  - type: table
-    name: Components & Features
-    filters:
-      or:
-        - file.hasTag("component")
-        - file.inFolder("features")
-    order:
-      - file.name
-      - title
-      - formula.doc_type
-      - formula.last_modified
-  - type: list
-    name: Recently Updated
-    order:
-      - file.name
-      - formula.last_modified
-    limit: 15
-  - type: table
-    name: Core Guides
-    filters:
-      not:
-        - file.inFolder("plugins")
-        - file.inFolder("features")
-        - file.inFolder("advanced")
-        - file.inFolder("tags")
-    order:
-      - file.name
-      - title
-      - formula.last_modified
--- a/docs/plugins/BasePage.md
+++ b/docs/plugins/BasePage.md
@@ -1,18 +0,0 @@
---
-title: BasePage
-tags:
-  - plugin/emitter
---
-
-This plugin emits pages for each view defined in `.base` files. See [[bases]] for usage.
-
-> [!note]
-> For information on how to add, remove or configure plugins, see the [[configuration#Plugins|Configuration]] page.
-
-Pages use `defaultListPageLayout` from `quartz.layout.ts` with `BaseContent` as the page body. To customize the layout, edit `quartz/components/pages/BaseContent.tsx`.
-
-## API
-
- Category: Emitter
- Function name: `Plugin.BasePage()`.
- Source: [`quartz/plugins/emitters/basePage.tsx`](https://github.com/jackyzha0/quartz/blob/v4/quartz/plugins/emitters/basePage.tsx).
--- a/docs/plugins/ObsidianBases.md
+++ b/docs/plugins/ObsidianBases.md
@@ -1,20 +0,0 @@
---
-title: ObsidianBases
-tags:
-  - plugin/transformer
---
-
-This plugin parses `.base` files and compiles them for rendering. See [[bases]] for usage.
-
-> [!note]
-> For information on how to add, remove or configure plugins, see the [[configuration#Plugins|Configuration]] page.
-
-## Configuration
-
- `emitWarnings`: If `true` (default), emits parse errors and type mismatches as warnings during build.
-
-## API
-
- Category: Transformer
- Function name: `Plugin.ObsidianBases()`.
- Source: [`quartz/plugins/transformers/bases.ts`](https://github.com/jackyzha0/quartz/blob/v4/quartz/plugins/transformers/bases.ts).
--- a/flake.lock
+++ b/flake.lock
@@ -0,0 +1,126 @@
+{
+  "nodes": {
+    "flake-utils": {
+      "inputs": {
+        "systems": "systems"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "flake-utils_2": {
+      "inputs": {
+        "systems": "systems_2"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1771008912,
+        "narHash": "sha256-gf2AmWVTs8lEq7z/3ZAsgnZDhWIckkb+ZnAo5RzSxJg=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "a82ccc39b39b621151d6732718e3e250109076fa",
+        "type": "github"
+      },
+      "original": {
+        "owner": "NixOS",
+        "ref": "nixos-unstable",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "nixpkgs_2": {
+      "locked": {
+        "lastModified": 1771369470,
+        "narHash": "sha256-0NBlEBKkN3lufyvFegY4TYv5mCNHbi5OmBDrzihbBMQ=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "0182a361324364ae3f436a63005877674cf45efb",
+        "type": "github"
+      },
+      "original": {
+        "owner": "NixOS",
+        "ref": "nixos-unstable",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "org-garden": {
+      "inputs": {
+        "flake-utils": "flake-utils_2",
+        "nixpkgs": "nixpkgs_2"
+      },
+      "locked": {
+        "path": "./org-garden",
+        "type": "path"
+      },
+      "original": {
+        "path": "./org-garden",
+        "type": "path"
+      },
+      "parent": []
+    },
+    "root": {
+      "inputs": {
+        "flake-utils": "flake-utils",
+        "nixpkgs": "nixpkgs",
+        "org-garden": "org-garden"
+      }
+    },
+    "systems": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    },
+    "systems_2": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}
--- a/flake.nix
+++ b/flake.nix
@@ -0,0 +1,47 @@
+{
+  description = "Quartz org-roam — org notes to website";
+
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
+    flake-utils.url = "github:numtide/flake-utils";
+    org-garden.url = "path:./org-garden";
+  };
+
+  outputs = { self, nixpkgs, flake-utils, org-garden }:
+    flake-utils.lib.eachDefaultSystem (system:
+      let
+        pkgs = import nixpkgs { inherit system; };
+
+        # Re-export org-garden's packages
+        orgGardenPkgs = org-garden.packages.${system};
+
+        # Convenience aliases
+        orgGardenApp = orgGardenPkgs.default;
+
+      in
+      {
+        # All packages come from org-garden
+        packages = orgGardenPkgs // {
+          default = orgGardenApp;
+        };
+
+        # Apps
+        apps = {
+          default = { type = "app"; program = "${orgGardenApp}/bin/org-garden"; };
+          org-garden = { type = "app"; program = "${orgGardenApp}/bin/org-garden"; };
+        };
+
+        # Dev shell for working on the repo
+        devShells.default = pkgs.mkShell {
+          buildInputs = [
+            pkgs.nodejs_22
+            pkgs.elixir
+          ];
+
+          shellHook = ''
+            echo "Node $(node --version) / npm $(npm --version)"
+            elixir --version 2>/dev/null | head -1 || true
+          '';
+        };
+      });
+}
--- a/notes-external/external-location-image.png
+++ b/notes-external/external-location-image.png
--- a/notes/bus/emt-madrid.org
+++ b/notes/bus/emt-madrid.org
@@ -0,0 +1,16 @@
+:PROPERTIES:
+:ID:       emt-madrid
+:END:
+#+title: EMT Madrid (urban bus)
+
+Empresa Municipal de Transportes (EMT) operates the urban bus network
+within the municipality of Madrid — around 200 lines.
+
+* Notable lines
+- *Line 27* — connects Embajadores with Barrio de la Concepción, one of the
+  oldest routes in the network.
+- *Line 34* — Argüelles to Carabanchel, crossing the city centre via Gran Vía.
+- *Búho (owl) lines* — night buses running from Cibeles from midnight to 6 am.
+
+* See also
+- [[id:madrid-transport][Madrid Public Transport]]
--- a/notes/example-citation.org
+++ b/notes/example-citation.org
@@ -0,0 +1,13 @@
+#+title: Example: Citation Reference
+
+This file demonstrates how org-citar citations pass through ox-hugo into
+markdown, where the pipeline transform resolves them.
+
+The methodology described in [cite:@podlovics2021journalArticle] provides a
+useful framework for analysis.
+
+Multiple citations can appear together:
+[cite:@podlovics2021journalArticle;@petersen2022book]
+
+Older bare-cite style (org-roam v1 / older citar) also works:
+cite:@podlovics2021journalArticle
--- a/notes/example-images.org
+++ b/notes/example-images.org
@@ -0,0 +1,33 @@
+:PROPERTIES:
+:ID:       example-images
+:END:
+#+title: Example: Image References
+
+This note demonstrates the three image reference scenarios that the pipeline
+must handle.
+
+* Scenario 1: External image (URL)
+
+An image hosted on the web — ox-hugo passes the URL through as-is and no
+local file handling is needed.
+
+#+attr_html: :link "https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSkzsTuLOt8esM6enoKwkzqA52G3p9hldlf2g&s"
+[[file:quartz-logo-external.png]]
+
+* Scenario 2: Local image (same notes directory)
+
+An image sitting next to this .org file inside the notes directory.
+ox-hugo copies files referenced with a relative path into the Hugo =static/=
+assets tree automatically.
+
+#+CAPTION: Quartz logo (local, same notes dir)
+[[file:quartz-logo.png]]
+
+* Scenario 3: External image (outside notes directory)
+
+An image that lives outside the notes directory entirely — for example a
+shared assets folder or a system path. ox-hugo still copies it into =static/=
+and rewrites the reference.
+
+#+CAPTION: Quartz logo (outside notes dir)
+[[file:../notes-external/external-location-image.png]]
--- a/notes/madrid-transport.org
+++ b/notes/madrid-transport.org
@@ -0,0 +1,17 @@
+:PROPERTIES:
+:ID:       madrid-transport
+:END:
+#+title: Madrid Public Transport
+
+Madrid has one of the most extensive public transport networks in Europe,
+operated primarily by [[id:crtm][Consorcio Regional de Transportes de Madrid]] (CRTM).
+
+* Modes
+- [[id:metro-madrid][Metro de Madrid]] — 13 lines, ~300 km of track
+- [[id:emt-madrid][EMT Bus]] — urban buses within the city
+- Cercanías — suburban rail run by Renfe
+- Interurbano — regional buses to the wider Community of Madrid
+
+* Ticketing
+A single [[https://www.crtm.es][tarjeta transporte]] (transport card) works across all modes.
+The Multi card covers zones A–C and is topped up at any metro station.
--- a/notes/metro/metro-madrid.org
+++ b/notes/metro/metro-madrid.org
@@ -0,0 +1,18 @@
+:PROPERTIES:
+:ID:       metro-madrid
+:END:
+#+title: Metro de Madrid
+
+The Madrid Metro is the main rapid transit network in the city, opened in 1919.
+It is the second oldest metro in the Iberian Peninsula after Barcelona.
+
+* Key Lines
+| Line | Name            | Colour | Terminals                    |
+|------+-----------------+--------+------------------------------|
+| L1   | Pinar de Chamartín–Valdecarros | Blue   | Pinar de Chamartín / Valdecarros |
+| L6   | Circular        | Grey   | Circular (loop)              |
+| L10  | —               | Dark blue | Hospital Infanta Sofía / Tres Olivos |
+
+* See also
+- [[id:madrid-transport][Madrid Public Transport]]
+- [[id:sol-interchange][Sol interchange]]
--- a/notes/metro/sol-interchange.org
+++ b/notes/metro/sol-interchange.org
@@ -0,0 +1,12 @@
+:PROPERTIES:
+:ID:       sol-interchange
+:END:
+#+title: Sol (interchange)
+
+Sol is the busiest interchange station in the Madrid Metro, sitting beneath
+Puerta del Sol in the city centre.
+
+Lines serving Sol: [[id:metro-madrid][L1]], L2, L3.
+
+It also connects to the Cercanías hub underneath, making it the de-facto
+zero point of Madrid's public transport.
--- a/notes/quartz-logo.png
+++ b/notes/quartz-logo.png
--- a/notes/roads/crtm.org
+++ b/notes/roads/crtm.org
@@ -0,0 +1,22 @@
+:PROPERTIES:
+:ID:       crtm
+:END:
+#+title: CRTM — Consorcio Regional de Transportes de Madrid
+
+The CRTM is the regional authority that coordinates public transport across
+the Community of Madrid. It does not operate services directly but sets
+fares, zones, and integration policy.
+
+* Fare zones
+| Zone  | Coverage                    |
+|-------+-----------------------------|
+| A     | Municipality of Madrid      |
+| B1    | Inner ring municipalities   |
+| B2    | Outer ring municipalities   |
+| B3    | Further suburban area       |
+| C1–C2 | Commuter belt               |
+
+* Related
+- [[id:madrid-transport][Madrid Public Transport]]
+- [[id:metro-madrid][Metro de Madrid]]
+- [[id:emt-madrid][EMT Madrid]]
--- a/notes/roads/m30.org
+++ b/notes/roads/m30.org
@@ -0,0 +1,19 @@
+:PROPERTIES:
+:ID:       m30
+:END:
+#+title: M-30
+
+The M-30 is Madrid's innermost ring road, circling the city centre at a
+radius of roughly 3–5 km from Puerta del Sol.
+
+It runs mostly underground through the Madrid Río tunnel section along the
+Manzanares river, built during the 2004–2007 renovation that reclaimed the
+riverbank as a public park.
+
+* Key junctions
+- Nudo Norte — connects to A-1 (Burgos) and A-6 (La Coruña)
+- Nudo Sur — connects to A-4 (Cádiz) and A-42 (Toledo)
+
+* See also
+- [[id:crtm][CRTM]]
+- [[id:madrid-transport][Madrid Public Transport]]
--- a/opencode.json
+++ b/opencode.json
@@ -0,0 +1,10 @@
+{
+  "$schema": "https://opencode.ai/config.json",
+  "mcp": {
+    "nixos": {
+      "type": "local",
+      "command": ["mcp-nixos"],
+      "enabled": true
+    }
+  }
+}
--- a/org-garden/flake.lock
+++ b/org-garden/flake.lock
@@ -0,0 +1,61 @@
+{
+  "nodes": {
+    "flake-utils": {
+      "inputs": {
+        "systems": "systems"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1771369470,
+        "narHash": "sha256-0NBlEBKkN3lufyvFegY4TYv5mCNHbi5OmBDrzihbBMQ=",
+        "owner": "NixOS",
+        "repo": "nixpkgs",
+        "rev": "0182a361324364ae3f436a63005877674cf45efb",
+        "type": "github"
+      },
+      "original": {
+        "owner": "NixOS",
+        "ref": "nixos-unstable",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "root": {
+      "inputs": {
+        "flake-utils": "flake-utils",
+        "nixpkgs": "nixpkgs"
+      }
+    },
+    "systems": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}
--- a/org-garden/flake.nix
+++ b/org-garden/flake.nix
@@ -0,0 +1,147 @@
+{
+  description = "Org-garden — org-roam to website publishing pipeline";
+
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
+    flake-utils.url = "github:numtide/flake-utils";
+  };
+
+  outputs = { self, nixpkgs, flake-utils }:
+    flake-utils.lib.eachDefaultSystem (system:
+      let
+        pkgs = import nixpkgs { inherit system; };
+        fs = pkgs.lib.fileset;
+
+        # =========================================================================
+        # Emacs with ox-hugo
+        # =========================================================================
+        # Needed at runtime by the escript (export calls `emacs --batch` with ox-hugo)
+        emacsWithOxHugo = (pkgs.emacsPackagesFor pkgs.emacs-nox).emacsWithPackages
+          (epkgs: [ epkgs.ox-hugo ]);
+
+        # =========================================================================
+        # Elixir Pipeline
+        # =========================================================================
+
+        # Pre-fetched Hex/Mix dependencies
+        mixDeps = pkgs.beamPackages.fetchMixDeps {
+          pname = "org-garden-mix-deps";
+          version = "0.1.0";
+          src = fs.toSource {
+            root = ./.;
+            fileset = fs.unions [
+              ./mix.exs
+              ./mix.lock
+            ];
+          };
+          sha256 = "sha256-si7JAomY1HZ33m6ihUJP5i6PO39CE1clYvuMtn0CbPU=";
+        };
+
+        # Compiled org-garden escript
+        orgGardenEscript = pkgs.beamPackages.mixRelease {
+          pname = "org-garden";
+          version = "0.1.0";
+          src = fs.toSource {
+            root = ./.;
+            fileset = fs.unions [
+              ./mix.exs
+              ./mix.lock
+              ./lib
+            ];
+          };
+          escriptBinName = "org_garden";
+          mixFodDeps = mixDeps;
+          stripDebug = true;
+        };
+
+        # =========================================================================
+        # Quartz (fetched from upstream, patched)
+        # =========================================================================
+
+        # Pin to specific upstream commit
+        quartzVersion = "4.5.2";
+        quartzRev = "ec00a40aefca73596ab76e3ebe3a8e1129b43688";
+
+        # Fetch upstream Quartz source
+        quartzSrc = pkgs.fetchFromGitHub {
+          owner = "jackyzha0";
+          repo = "quartz";
+          rev = quartzRev;
+          hash = "sha256-HdtQB5+SRWiypOvAJuJa3Nodl4JHehp2Mz6Rj5gOG0w=";
+        };
+
+        # Apply our patches to Quartz
+        quartzPatched = pkgs.runCommand "quartz-patched-${quartzVersion}" {
+          src = quartzSrc;
+        } ''
+          cp -r $src $out
+          chmod -R u+w $out
+          cd $out
+          patch -p1 < ${./patches/01-glob-gitignore.patch}
+          patch -p1 < ${./patches/02-build-gitignore.patch}
+          patch -p1 < ${./patches/03-static-hugo.patch}
+          patch -p1 < ${./patches/04-oxhugofm-figure.patch}
+        '';
+
+        # Pre-fetch Quartz npm dependencies
+        quartzDeps = pkgs.buildNpmPackage {
+          pname = "org-garden-quartz-deps";
+          version = quartzVersion;
+          src = quartzPatched;
+          npmDepsHash = "sha256-7u+VlIx44B3/ivM9vLMIOn+e4TL4eS6B682vhS+Ikb4=";
+          dontBuild = true;
+          installPhase = ''
+            mkdir -p $out
+            cp -r node_modules $out/node_modules
+          '';
+        };
+
+        # =========================================================================
+        # Combined Application
+        # =========================================================================
+
+        # Wrapped org-garden with Quartz bundled
+        orgGardenApp = pkgs.writeShellApplication {
+          name = "org-garden";
+          runtimeInputs = [ emacsWithOxHugo pkgs.inotify-tools pkgs.nodejs_22 ];
+          text = ''
+            # Set up Quartz working directory
+            QUARTZ_WORK=$(mktemp -d)
+            trap 'rm -rf "$QUARTZ_WORK"' EXIT
+
+            # Copy patched Quartz source
+            cp -r ${quartzPatched}/. "$QUARTZ_WORK/"
+            chmod -R u+w "$QUARTZ_WORK"
+
+            # Copy default config files
+            cp ${./quartz-config/quartz.config.ts} "$QUARTZ_WORK/"
+            cp ${./quartz-config/quartz.layout.ts} "$QUARTZ_WORK/"
+            cp ${./quartz-config/globals.d.ts} "$QUARTZ_WORK/"
+            cp ${./quartz-config/index.d.ts} "$QUARTZ_WORK/"
+
+            # Link pre-built node_modules
+            ln -s ${quartzDeps}/node_modules "$QUARTZ_WORK/node_modules"
+
+            export QUARTZ_PATH="$QUARTZ_WORK"
+            export NODE_PATH="${pkgs.nodejs_22}/bin/node"
+
+            exec ${orgGardenEscript}/bin/org_garden "$@"
+          '';
+        };
+
+      in
+      {
+        packages.default = orgGardenApp;
+        packages.escript = orgGardenEscript;
+        packages.quartz-patched = quartzPatched;
+
+        devShells.default = pkgs.mkShell {
+          buildInputs = [
+            pkgs.elixir
+            pkgs.inotify-tools
+            emacsWithOxHugo
+            pkgs.nodejs_22
+          ];
+        };
+      });
+}
--- a/org-garden/lib/org_garden.ex
+++ b/org-garden/lib/org_garden.ex
@@ -0,0 +1,189 @@
+defmodule OrgGarden do
+  @moduledoc """
+  Org-roam to website publishing pipeline.
+
+  Orchestrates:
+  1. Org → Markdown export (via Emacs + ox-hugo)
+  2. Markdown transforms (citations, etc.)
+  3. Markdown → HTML + serving (via Quartz)
+
+  ## Usage
+
+      opts = %{
+        zotero_url: "http://localhost:23119",
+        bibtex_file: System.get_env("BIBTEX_FILE"),
+        citation_mode: :warn   # :silent | :warn | :strict
+      }
+
+      # Batch: all .md files in a directory
+      OrgGarden.run(content_dir, [OrgGarden.Transforms.Citations], opts)
+
+      # Targeted: specific files only
+      OrgGarden.run_on_files(["content/foo.md"], [OrgGarden.Transforms.Citations], opts)
+
+      # With pre-initialized transforms (for watch mode, avoids re-init)
+      initialized = OrgGarden.init_transforms([OrgGarden.Transforms.Citations], opts)
+      OrgGarden.run_on_files_with(["content/foo.md"], initialized, opts)
+  """
+
+  require Logger
+
+  @type transform :: module()
+  @type initialized_transform :: {module(), term()}
+  @type opts :: map()
+
+  @doc "One-shot build: org files → static site"
+  def build(notes_dir, opts \\ []) do
+    OrgGarden.CLI.handle_build([notes_dir | opts_to_args(opts)])
+  end
+
+  @doc "Development server: watch + live reload"
+  def serve(notes_dir, opts \\ []) do
+    OrgGarden.CLI.handle_serve([notes_dir | opts_to_args(opts)])
+  end
+
+  @doc "Export only: org files → markdown (no Quartz)"
+  def export(notes_dir, opts \\ []) do
+    OrgGarden.CLI.handle_export([notes_dir | opts_to_args(opts)])
+  end
+
+  defp opts_to_args(opts) do
+    Enum.flat_map(opts, fn
+      {:output, v} -> ["--output", v]
+      {:port, v} -> ["--port", to_string(v)]
+      {:ws_port, v} -> ["--ws-port", to_string(v)]
+      {:watch, true} -> ["--watch"]
+      {:watch, false} -> []
+      _ -> []
+    end)
+  end
+
+  @doc """
+  Initialize transform modules. Returns a list of `{module, state}` tuples.
+
+  Call this once and reuse the result with `run_on_files_with/3` to avoid
+  re-initializing transforms on every file change (e.g., in watch mode).
+  """
+  @spec init_transforms([transform()], opts()) :: [initialized_transform()]
+  def init_transforms(transforms, opts) do
+    Enum.map(transforms, fn mod ->
+      state = mod.init(opts)
+      {mod, state}
+    end)
+  end
+
+  @doc """
+  Tear down previously initialized transforms, releasing any resources.
+  """
+  @spec teardown_transforms([initialized_transform()]) :: :ok
+  def teardown_transforms(initialized) do
+    Enum.each(initialized, fn {mod, state} ->
+      if function_exported?(mod, :teardown, 1) do
+        mod.teardown(state)
+      end
+    end)
+
+    :ok
+  end
+
+  @doc """
+  Run all transforms over every `.md` file under `content_dir`.
+
+  Initializes and tears down transforms automatically.
+  Returns `{:ok, stats}` where stats maps each transform to a count of files it changed.
+  """
+  @spec run(String.t(), [transform()], opts()) :: {:ok, map()}
+  def run(content_dir, transforms, opts \\ %{}) do
+    md_files =
+      content_dir
+      |> Path.join("**/*.md")
+      |> Path.wildcard()
+
+    if md_files == [] do
+      Logger.warning("OrgGarden: no .md files found in #{content_dir}")
+      {:ok, %{}}
+    else
+      Logger.info(
+        "OrgGarden: processing #{length(md_files)} markdown files " <>
+          "with #{length(transforms)} transform(s)"
+      )
+
+      initialized = init_transforms(transforms, opts)
+      stats = apply_transforms(md_files, initialized, opts)
+      teardown_transforms(initialized)
+      {:ok, stats}
+    end
+  end
+
+  @doc """
+  Run all transforms over specific `.md` files only.
+
+  Initializes and tears down transforms automatically.
+  Files that don't exist are silently skipped.
+  """
+  @spec run_on_files([String.t()], [transform()], opts()) :: {:ok, map()}
+  def run_on_files(file_paths, transforms, opts \\ %{}) do
+    existing = Enum.filter(file_paths, &File.exists?/1)
+
+    if existing == [] do
+      Logger.debug("OrgGarden: no files to process")
+      {:ok, %{}}
+    else
+      Logger.info("OrgGarden: processing #{length(existing)} file(s)")
+      initialized = init_transforms(transforms, opts)
+      stats = apply_transforms(existing, initialized, opts)
+      teardown_transforms(initialized)
+      {:ok, stats}
+    end
+  end
+
+  @doc """
+  Run pre-initialized transforms over specific `.md` files.
+
+  Does NOT call `init` or `teardown` — the caller manages the transform
+  lifecycle. Use this in watch mode to avoid re-initializing on every change.
+  """
+  @spec run_on_files_with([String.t()], [initialized_transform()], opts()) :: {:ok, map()}
+  def run_on_files_with(file_paths, initialized, opts) do
+    existing = Enum.filter(file_paths, &File.exists?/1)
+
+    if existing == [] do
+      Logger.debug("OrgGarden: no files to process")
+      {:ok, %{}}
+    else
+      stats = apply_transforms(existing, initialized, opts)
+      {:ok, stats}
+    end
+  end
+
+  # -------------------------------------------------------------------
+  # Private
+  # -------------------------------------------------------------------
+
+  defp apply_transforms(md_files, initialized, opts) do
+    Enum.reduce(md_files, %{}, fn path, acc ->
+      original = File.read!(path)
+
+      {transformed, file_stats} =
+        Enum.reduce(initialized, {original, %{}}, fn {mod, state}, {content, fstats} ->
+          result = mod.apply(content, state, opts)
+          changed = result != content
+
+          {result,
+           Map.update(
+             fstats,
+             mod,
+             if(changed, do: 1, else: 0),
+             &(&1 + if(changed, do: 1, else: 0))
+           )}
+        end)
+
+      if transformed != original do
+        File.write!(path, transformed)
+        Logger.debug("OrgGarden: updated #{Path.relative_to_cwd(path)}")
+      end
+
+      Map.merge(acc, file_stats, fn _k, a, b -> a + b end)
+    end)
+  end
+end
--- a/org-garden/lib/org_garden/application.ex
+++ b/org-garden/lib/org_garden/application.ex
@@ -0,0 +1,14 @@
+defmodule OrgGarden.Application do
+  @moduledoc false
+  use Application
+
+  @impl true
+  def start(_type, _args) do
+    children = [
+      {Finch, name: OrgGarden.Finch}
+    ]
+
+    opts = [strategy: :one_for_one, name: OrgGarden.AppSupervisor]
+    Supervisor.start_link(children, opts)
+  end
+end
--- a/org-garden/lib/org_garden/cli.ex
+++ b/org-garden/lib/org_garden/cli.ex
@@ -0,0 +1,375 @@
+defmodule OrgGarden.CLI do
+  @moduledoc """
+  Escript entry point for the org-garden pipeline.
+
+  ## Commands
+
+      org-garden serve <notes-dir> [--port 8080] [--ws-port 3001]
+      org-garden build <notes-dir> [--output <path>]
+      org-garden export <notes-dir> [--watch]
+
+  ### serve
+  Development server with watch + live reload. Starts both the org→md
+  watcher and Quartz in serve mode.
+
+  ### build
+  One-shot build for CI/production. Exports org files, runs transforms,
+  then builds static site with Quartz.
+
+  ### export
+  Just export org→md (current pipeline behavior). Use --watch for
+  incremental re-export on file changes.
+
+  ## Arguments
+
+    notes-dir          Path to the directory containing `.org` notes (required).
+                       Also accepts the `NOTES_DIR` env var.
+
+  ## Options
+
+    --output <path>    Output root directory (used as ox-hugo base dir).
+                       Defaults to the `OUTPUT_DIR` env var, or the current
+                       working directory.
+    --content-dir <p>  Output directory for exported Markdown. Defaults to
+                       `<output>/content`.
+    --port <n>         HTTP server port (default: 8080). Only for `serve`.
+    --ws-port <n>      WebSocket hot reload port (default: 3001). Only for `serve`.
+    --watch            After initial batch, watch notes-dir for changes and
+                       incrementally re-export affected files. Only for `export`.
+
+  ## Environment Variables
+
+    BIBTEX_FILE        Path to a `.bib` file used as citation fallback.
+    ZOTERO_URL         Zotero Better BibTeX base URL (default: http://localhost:23119).
+    CITATION_MODE      silent | warn (default) | strict.
+    QUARTZ_PATH        Path to quartz directory (required for serve/build).
+    NODE_PATH          Node.js executable (default: node).
+  """
+
+  require Logger
+
+  @transforms [OrgGarden.Transforms.Citations]
+
+  def main(argv) do
+    Application.ensure_all_started(:org_garden)
+
+    case argv do
+      ["serve" | rest] -> handle_serve(rest)
+      ["build" | rest] -> handle_build(rest)
+      ["export" | rest] -> handle_export(rest)
+      # Legacy: treat bare args as export command for backward compatibility
+      [_ | _] -> handle_export(argv)
+      _ -> abort("Usage: org-garden <serve|build|export> <notes-dir> [options]")
+    end
+  end
+
+  # ---------------------------------------------------------------------------
+  # Command: serve
+  # ---------------------------------------------------------------------------
+
+  def handle_serve(argv) do
+    require_quartz_env()
+    {notes_dir, output_dir, content_dir, opts} = parse_serve_args(argv)
+    pipeline_opts = build_pipeline_opts()
+
+    # Initial batch export
+    wipe(content_dir)
+    export_all(notes_dir, output_dir)
+    run_pipeline(content_dir, pipeline_opts)
+    generate_index(content_dir)
+
+    IO.puts("==> Starting development server...")
+
+    {:ok, _pid} =
+      OrgGarden.Supervisor.start_link(
+        notes_dir: notes_dir,
+        output_dir: output_dir,
+        content_dir: content_dir,
+        pipeline_opts: pipeline_opts,
+        transforms: @transforms,
+        port: opts[:port] || 8080,
+        ws_port: opts[:ws_port] || 3001
+      )
+
+    IO.puts("==> Server running at http://localhost:#{opts[:port] || 8080}")
+    IO.puts("==> Watching #{notes_dir} for changes (Ctrl+C to stop)")
+
+    Process.sleep(:infinity)
+  end
+
+  defp parse_serve_args(argv) do
+    {opts, positional, _invalid} =
+      OptionParser.parse(argv,
+        strict: [
+          output: :string,
+          content_dir: :string,
+          port: :integer,
+          ws_port: :integer
+        ]
+      )
+
+    notes_dir = extract_notes_dir(positional, "serve")
+    output_dir = extract_output_dir(opts)
+    content_dir = extract_content_dir(opts, output_dir)
+
+    {notes_dir, output_dir, content_dir, opts}
+  end
+
+  # ---------------------------------------------------------------------------
+  # Command: build
+  # ---------------------------------------------------------------------------
+
+  def handle_build(argv) do
+    quartz_path = require_quartz_env()
+    {notes_dir, output_dir, content_dir, _opts} = parse_build_args(argv)
+    pipeline_opts = build_pipeline_opts()
+
+    # Full batch export
+    wipe(content_dir)
+    export_all(notes_dir, output_dir)
+    run_pipeline(content_dir, pipeline_opts)
+    generate_index(content_dir)
+
+    node_path = System.get_env("NODE_PATH", "node")
+
+    IO.puts("==> Building static site with Quartz...")
+
+    {output, status} =
+      System.cmd(
+        node_path,
+        [
+          Path.join(quartz_path, "quartz/bootstrap-cli.mjs"),
+          "build",
+          "--directory",
+          content_dir,
+          "--output",
+          Path.join(output_dir, "public")
+        ],
+        cd: quartz_path,
+        stderr_to_stdout: true
+      )
+
+    IO.puts(output)
+
+    if status != 0 do
+      abort("Quartz build failed with status #{status}")
+    end
+
+    IO.puts("==> Build complete. Output: #{Path.join(output_dir, "public")}")
+  end
+
+  defp parse_build_args(argv) do
+    {opts, positional, _invalid} =
+      OptionParser.parse(argv,
+        strict: [output: :string, content_dir: :string]
+      )
+
+    notes_dir = extract_notes_dir(positional, "build")
+    output_dir = extract_output_dir(opts)
+    content_dir = extract_content_dir(opts, output_dir)
+
+    {notes_dir, output_dir, content_dir, opts}
+  end
+
+  # ---------------------------------------------------------------------------
+  # Command: export (original pipeline behavior)
+  # ---------------------------------------------------------------------------
+
+  def handle_export(argv) do
+    {notes_dir, output_dir, content_dir, watch?} = parse_export_args(argv)
+    pipeline_opts = build_pipeline_opts()
+
+    # Phase 1-4: full batch export
+    wipe(content_dir)
+    export_all(notes_dir, output_dir)
+    run_pipeline(content_dir, pipeline_opts)
+    generate_index(content_dir)
+
+    md_count =
+      content_dir
+      |> Path.join("**/*.md")
+      |> Path.wildcard()
+      |> length()
+
+    IO.puts("==> Done. #{md_count} markdown files in #{content_dir}")
+
+    # Phase 5: optional watch mode
+    if watch? do
+      IO.puts("==> Watching #{notes_dir} for .org changes... (Ctrl+C to stop)")
+
+      {:ok, _pid} =
+        OrgGarden.Watcher.start_link(
+          notes_dir: notes_dir,
+          output_dir: output_dir,
+          content_dir: content_dir,
+          pipeline_opts: pipeline_opts,
+          transforms: @transforms
+        )
+
+      Process.sleep(:infinity)
+    end
+  end
+
+  defp parse_export_args(argv) do
+    {opts, positional, _invalid} =
+      OptionParser.parse(argv,
+        strict: [output: :string, content_dir: :string, watch: :boolean]
+      )
+
+    notes_dir = extract_notes_dir(positional, "export")
+    output_dir = extract_output_dir(opts)
+    content_dir = extract_content_dir(opts, output_dir)
+    watch? = Keyword.get(opts, :watch, false)
+
+    {notes_dir, output_dir, content_dir, watch?}
+  end
+
+  # ---------------------------------------------------------------------------
+  # Shared argument extraction
+  # ---------------------------------------------------------------------------
+
+  defp extract_notes_dir(positional, command) do
+    notes_dir =
+      case positional do
+        [dir | _] ->
+          dir
+
+        [] ->
+          System.get_env("NOTES_DIR") ||
+            abort("Usage: org-garden #{command} <notes-dir> [options]")
+      end
+
+    notes_dir = Path.expand(notes_dir)
+
+    unless File.dir?(notes_dir) do
+      abort("Error: notes directory does not exist: #{notes_dir}")
+    end
+
+    notes_dir
+  end
+
+  defp extract_output_dir(opts) do
+    (opts[:output] || System.get_env("OUTPUT_DIR") || File.cwd!())
+    |> Path.expand()
+  end
+
+  defp extract_content_dir(opts, output_dir) do
+    (opts[:content_dir] || Path.join(output_dir, "content"))
+    |> Path.expand()
+  end
+
+  # ---------------------------------------------------------------------------
+  # Phase 1: Wipe content/
+  # ---------------------------------------------------------------------------
+
+  defp wipe(content_dir) do
+    IO.puts("==> Wiping #{content_dir}")
+    File.mkdir_p!(content_dir)
+
+    content_dir
+    |> File.ls!()
+    |> Enum.reject(&(&1 == ".gitkeep"))
+    |> Enum.each(fn entry ->
+      Path.join(content_dir, entry) |> File.rm_rf!()
+    end)
+  end
+
+  # ---------------------------------------------------------------------------
+  # Phase 2: Export org files via Emacs + ox-hugo
+  # ---------------------------------------------------------------------------
+
+  defp export_all(notes_dir, output_dir) do
+    IO.puts("==> Exporting org files from #{notes_dir}")
+
+    case OrgGarden.Export.export_all(notes_dir, output_dir) do
+      {:ok, 0} ->
+        IO.puts("No .org files found in #{notes_dir}")
+        System.halt(0)
+
+      {:ok, count} ->
+        IO.puts("  exported #{count} file(s)")
+
+      {:error, failures} ->
+        IO.puts(:stderr, "\nFailed to export #{length(failures)} file(s):")
+
+        Enum.each(failures, fn {f, {:error, reason}} ->
+          IO.puts(:stderr, "  #{f}: #{inspect(reason)}")
+        end)
+
+        System.halt(1)
+    end
+  end
+
+  # ---------------------------------------------------------------------------
+  # Phase 3: Markdown transformation pipeline
+  # ---------------------------------------------------------------------------
+
+  defp run_pipeline(content_dir, pipeline_opts) do
+    IO.puts("==> Running markdown pipeline")
+
+    {:ok, stats} = OrgGarden.run(content_dir, @transforms, pipeline_opts)
+
+    Enum.each(stats, fn {mod, count} ->
+      IO.puts("  #{inspect(mod)}: #{count} file(s) modified")
+    end)
+  end
+
+  # ---------------------------------------------------------------------------
+  # Phase 4: Generate default index.md if none was exported
+  # ---------------------------------------------------------------------------
+
+  defp generate_index(content_dir) do
+    IO.puts("==> Generating index")
+    OrgGarden.Index.generate(content_dir)
+  end
+
+  # ---------------------------------------------------------------------------
+  # Helpers
+  # ---------------------------------------------------------------------------
+
+  defp require_quartz_env do
+    case System.get_env("QUARTZ_PATH") do
+      nil ->
+        abort("""
+        Error: QUARTZ_PATH environment variable not set.
+
+        The 'serve' and 'build' commands require Quartz to be available.
+
+        Use the wrapper scripts that set up the environment:
+          nix run .#notes -- <notes-dir>     # for serve
+          nix run .#build -- <notes-dir>     # for build
+
+        Or set QUARTZ_PATH manually to point to a quartz-org-roam checkout
+        with node_modules installed.
+
+        For export-only mode (no Quartz), use:
+          org-garden export <notes-dir> [--watch]
+        """)
+
+      path ->
+        unless File.exists?(Path.join(path, "quartz/bootstrap-cli.mjs")) do
+          abort("Error: QUARTZ_PATH=#{path} does not contain quartz/bootstrap-cli.mjs")
+        end
+
+        path
+    end
+  end
+
+  defp build_pipeline_opts do
+    %{
+      zotero_url: System.get_env("ZOTERO_URL", "http://localhost:23119"),
+      bibtex_file: System.get_env("BIBTEX_FILE"),
+      citation_mode:
+        case System.get_env("CITATION_MODE", "warn") do
+          "silent" -> :silent
+          "strict" -> :strict
+          _ -> :warn
+        end
+    }
+  end
+
+  defp abort(message) do
+    IO.puts(:stderr, message)
+    System.halt(1)
+  end
+end
--- a/org-garden/lib/org_garden/export.ex
+++ b/org-garden/lib/org_garden/export.ex
@@ -0,0 +1,135 @@
+defmodule OrgGarden.Export do
+  @moduledoc """
+  Org-to-Markdown export via Emacs batch + ox-hugo.
+
+  Provides both single-file and batch export, plus a helper to compute
+  the expected `.md` output path for a given `.org` source file.
+  """
+
+  require Logger
+
+  @doc """
+  Export a single `.org` file to Markdown via `emacs --batch` + ox-hugo.
+
+  Returns `{:ok, exit_code}` with the emacs exit code (0 = success),
+  or `{:error, reason}` if the command could not be executed.
+  """
+  @spec export_file(String.t(), String.t(), String.t()) :: {:ok, non_neg_integer()} | {:error, term()}
+  def export_file(orgfile, notes_dir, output_dir) do
+    section =
+      orgfile
+      |> Path.dirname()
+      |> Path.relative_to(notes_dir)
+
+    # ox-hugo requires static/ to exist for image asset copying
+    File.mkdir_p!(Path.join(output_dir, "static"))
+
+    {output, exit_code} =
+      System.cmd(
+        "emacs",
+        [
+          "--batch",
+          "--eval", "(require 'ox-hugo)",
+          "--eval", """
+          (org-cite-register-processor 'passthrough
+            :export-citation
+            (lambda (citation _style _backend _info)
+              (let ((keys (mapcar (lambda (ref)
+                                    (concat "@" (org-element-property :key ref)))
+                                  (org-cite-get-references citation))))
+                (format "[cite:%s]" (string-join keys ";")))))
+          """,
+          "--eval", "(setq org-cite-export-processors '((t passthrough)))",
+          "--eval", ~s[(setq org-hugo-base-dir "#{output_dir}")],
+          "--eval", ~s[(setq org-hugo-default-section-directory "#{section}")],
+          "--visit", orgfile,
+          "--funcall", "org-hugo-export-to-md"
+        ],
+        stderr_to_stdout: true
+      )
+
+    filtered =
+      output
+      |> String.split("\n")
+      |> Enum.reject(&String.match?(&1, ~r/^Loading|^ad-handle|^For information/))
+      |> Enum.join("\n")
+
+    if filtered != "", do: Logger.info("emacs: #{filtered}")
+
+    if exit_code == 0 do
+      {:ok, exit_code}
+    else
+      {:error, {:emacs_exit, exit_code, filtered}}
+    end
+  rescue
+    e -> {:error, e}
+  end
+
+  @doc """
+  Export all `.org` files found under `notes_dir`.
+
+  Returns `{:ok, count}` where `count` is the number of successfully
+  exported files, or `{:error, failures}` if any files failed.
+  """
+  @spec export_all(String.t(), String.t()) :: {:ok, non_neg_integer()} | {:error, list()}
+  def export_all(notes_dir, output_dir) do
+    org_files =
+      Path.join(notes_dir, "**/*.org")
+      |> Path.wildcard()
+
+    if org_files == [] do
+      Logger.warning("No .org files found in #{notes_dir}")
+      {:ok, 0}
+    else
+      Logger.info("Exporting #{length(org_files)} org file(s) from #{notes_dir}")
+
+      results =
+        Enum.map(org_files, fn orgfile ->
+          IO.puts("  exporting: #{orgfile}")
+          {orgfile, export_file(orgfile, notes_dir, output_dir)}
+        end)
+
+      failures =
+        Enum.filter(results, fn
+          {_, {:ok, _}} -> false
+          {_, {:error, _}} -> true
+        end)
+
+      if failures == [] do
+        {:ok, length(results)}
+      else
+        {:error, failures}
+      end
+    end
+  end
+
+  @doc """
+  Compute the expected `.md` path for a given `.org` file.
+
+  Uses the same section-mapping logic as ox-hugo: the relative directory
+  of the `.org` file within `notes_dir` becomes the section directory
+  under `content_dir`.
+
+  ## Examples
+
+      iex> OrgGarden.Export.expected_md_path("/notes/bus/emt.org", "/notes", "/out/content")
+      "/out/content/bus/emt.md"
+
+      iex> OrgGarden.Export.expected_md_path("/notes/top-level.org", "/notes", "/out/content")
+      "/out/content/top-level.md"
+  """
+  @spec expected_md_path(String.t(), String.t(), String.t()) :: String.t()
+  def expected_md_path(orgfile, notes_dir, content_dir) do
+    section =
+      orgfile
+      |> Path.dirname()
+      |> Path.relative_to(notes_dir)
+
+    basename = Path.basename(orgfile, ".org") <> ".md"
+
+    case section do
+      "." -> Path.join(content_dir, basename)
+      _ -> Path.join([content_dir, section, basename])
+    end
+  end
+end
--- a/org-garden/lib/org_garden/index.ex
+++ b/org-garden/lib/org_garden/index.ex
@@ -0,0 +1,83 @@
+defmodule OrgGarden.Index do
+  @moduledoc """
+  Generates a fallback `index.md` in the content directory if none was
+  exported from an `.org` file.
+
+  The generated index lists all markdown pages alphabetically with links.
+  """
+
+  @doc """
+  Generate `content_dir/index.md` if it does not already exist.
+
+  If an `index.md` was already created by ox-hugo (from an `index.org`),
+  it is left untouched.
+  """
+  @spec generate(String.t()) :: :ok
+  def generate(content_dir) do
+    index_path = Path.join(content_dir, "index.md")
+
+    unless File.exists?(index_path) do
+      IO.puts("  generating default index.md")
+
+      pages =
+        Path.join(content_dir, "**/*.md")
+        |> Path.wildcard()
+        |> Enum.map(fn path ->
+          slug = Path.relative_to(path, content_dir) |> Path.rootname()
+
+          title =
+            path
+            |> File.read!()
+            |> then(fn content ->
+              case Regex.run(~r/^title\s*=\s*"(.+)"/m, content) do
+                [_, t] -> t
+                _ -> slug
+              end
+            end)
+
+          {slug, title}
+        end)
+        |> Enum.sort_by(fn {_, title} -> title end)
+        |> Enum.map(fn {slug, title} -> "- [#{title}](#{slug})" end)
+        |> Enum.join("\n")
+
+      File.write!(index_path, """
+      ---
+      title: Index
+      ---
+
+      #{pages}
+      """)
+    end
+
+    :ok
+  end
+
+  @doc """
+  Regenerate the index by removing any previously generated one first.
+
+  Only removes the index if it was generated by us (contains `title: Index`).
+  User-exported index files (from `index.org`) are left untouched.
+  """
+  @spec regenerate(String.t()) :: :ok
+  def regenerate(content_dir) do
+    index_path = Path.join(content_dir, "index.md")
+
+    if File.exists?(index_path) do
+      content = File.read!(index_path)
+
+      if generated_index?(content) do
+        File.rm!(index_path)
+      end
+    end
+
+    generate(content_dir)
+  end
+
+  defp generated_index?(content) do
+    # Our generated index uses "title: Index" in YAML frontmatter.
+    # ox-hugo uses TOML frontmatter (title = "..."), so this won't
+    # match user-exported files.
+    String.contains?(content, "title: Index")
+  end
+end
--- a/org-garden/lib/org_garden/quartz.ex
+++ b/org-garden/lib/org_garden/quartz.ex
@@ -0,0 +1,118 @@
+defmodule OrgGarden.Quartz do
+  @moduledoc """
+  Manages Quartz Node.js process as an Erlang Port.
+
+  Required environment:
+  - QUARTZ_PATH: path to quartz repo (with node_modules)
+  - NODE_PATH: path to node executable (default: "node")
+
+  Starts Quartz in serve mode (`npx quartz build --serve`) and forwards
+  all stdout/stderr output to the Logger with a `[quartz]` prefix.
+
+  If Quartz exits, this GenServer will stop, which triggers the supervisor
+  to restart the entire supervision tree (strategy: :one_for_all).
+  """
+  use GenServer
+
+  require Logger
+
+  defstruct [:port, :quartz_path, :content_dir, :http_port, :ws_port]
+
+  # -------------------------------------------------------------------
+  # Client API
+  # -------------------------------------------------------------------
+
+  @doc """
+  Start the Quartz process as a linked GenServer.
+
+  ## Options
+
+    * `:content_dir` — directory where markdown files are located (required)
+    * `:port`        — HTTP server port (default: 8080)
+    * `:ws_port`     — WebSocket hot reload port (default: 3001)
+  """
+  def start_link(opts) do
+    GenServer.start_link(__MODULE__, opts, name: __MODULE__)
+  end
+
+  # -------------------------------------------------------------------
+  # GenServer callbacks
+  # -------------------------------------------------------------------
+
+  @impl true
+  def init(opts) do
+    quartz_path =
+      System.get_env("QUARTZ_PATH") ||
+        raise "QUARTZ_PATH environment variable not set"
+
+    node_path = System.get_env("NODE_PATH", "node")
+
+    content_dir = Keyword.fetch!(opts, :content_dir)
+    http_port = Keyword.get(opts, :port, 8080)
+    ws_port = Keyword.get(opts, :ws_port, 3001)
+
+    cli_path = Path.join(quartz_path, "quartz/bootstrap-cli.mjs")
+
+    unless File.exists?(cli_path) do
+      raise "Quartz CLI not found at #{cli_path}. Check QUARTZ_PATH."
+    end
+
+    args = [
+      cli_path,
+      "build",
+      "--serve",
+      "--directory", content_dir,
+      "--port", to_string(http_port),
+      "--wsPort", to_string(ws_port)
+    ]
+
+    Logger.info("[quartz] Starting: #{node_path} #{Enum.join(args, " ")}")
+    Logger.info("[quartz] Working directory: #{quartz_path}")
+
+    port =
+      Port.open({:spawn_executable, node_path}, [
+        :binary,
+        :exit_status,
+        :stderr_to_stdout,
+        args: args,
+        cd: quartz_path,
+        env: [{~c"NODE_NO_WARNINGS", ~c"1"}]
+      ])
+
+    state = %__MODULE__{
+      port: port,
+      quartz_path: quartz_path,
+      content_dir: content_dir,
+      http_port: http_port,
+      ws_port: ws_port
+    }
+
+    {:ok, state}
+  end
+
+  @impl true
+  def handle_info({port, {:data, data}}, %{port: port} = state) do
+    data
+    |> String.split("\n", trim: true)
+    |> Enum.each(&Logger.info("[quartz] #{&1}"))
+
+    {:noreply, state}
+  end
+
+  @impl true
+  def handle_info({port, {:exit_status, status}}, %{port: port} = state) do
+    Logger.error("[quartz] Process exited with status #{status}")
+    {:stop, {:quartz_exit, status}, state}
+  end
+
+  @impl true
+  def terminate(_reason, %{port: port}) when is_port(port) do
+    # Attempt graceful shutdown
+    Port.close(port)
+    :ok
+  rescue
+    _ -> :ok
+  end
+
+  def terminate(_reason, _state), do: :ok
+end
--- a/org-garden/lib/org_garden/resolvers/bibtex.ex
+++ b/org-garden/lib/org_garden/resolvers/bibtex.ex
@@ -0,0 +1,178 @@
+defmodule OrgGarden.Resolvers.BibTeX do
+  @moduledoc """
+  Resolves citation keys from a local BibTeX (.bib) file.
+
+  Configured via the `BIBTEX_FILE` environment variable, or passed directly
+  as `opts.bibtex_file`. The file is parsed once at init time and the
+  resulting entry map is reused for all lookups.
+
+  Supports extracting: author last names, year, title, DOI, URL.
+
+  BibTeX entry format parsed:
+
+      @type{citationkey,
+        author  = {Last, First and Last2, First2},
+        year    = {2021},
+        title   = {Some Title},
+        doi     = {10.xxxx/yyyy},
+        url     = {https://example.com},
+      }
+
+  Returns `{:ok, %{label: "Author, Year", url: "..."}}` or `:error`.
+  """
+
+  require Logger
+
+  # ------------------------------------------------------------------
+  # Public API
+  # ------------------------------------------------------------------
+
+  @doc """
+  Parse a .bib file and return a map of `%{citation_key => entry_map}`.
+  Returns `{:ok, entries}` or `{:error, reason}`.
+  """
+  @spec load(String.t()) :: {:ok, map()} | {:error, term()}
+  def load(path) do
+    case File.read(path) do
+      {:ok, content} ->
+        entries = parse_entries(content)
+        Logger.info("BibTeX: loaded #{map_size(entries)} entries from #{path}")
+        {:ok, entries}
+
+      {:error, reason} ->
+        {:error, reason}
+    end
+  end
+
+  @doc """
+  Resolve a citation key from pre-loaded BibTeX entries.
+  """
+  @spec resolve(String.t(), map()) :: {:ok, map()} | :error
+  def resolve(key, entries) do
+    case Map.fetch(entries, key) do
+      {:ok, entry} ->
+        label = build_label(entry)
+        url = build_url(entry)
+        {:ok, %{label: label, url: url}}
+
+      :error ->
+        :error
+    end
+  end
+
+  # ------------------------------------------------------------------
+  # Parsing
+  # ------------------------------------------------------------------
+
+  # Match @type{key, ...fields...}
+  # We handle nested braces by scanning character by character after
+  # finding the opening, rather than relying on a single regex.
+  @entry_header ~r/@\w+\s*\{\s*([^,\s]+)\s*,/
+
+  defp parse_entries(content) do
+    # Split on "@" boundaries, then parse each chunk
+    content
+    |> String.split(~r/(?=@\w+\s*\{)/, trim: true)
+    |> Enum.reduce(%{}, fn chunk, acc ->
+      case Regex.run(@entry_header, chunk) do
+        [_, key] ->
+          fields = parse_fields(chunk)
+          Map.put(acc, String.trim(key), fields)
+
+        _ ->
+          acc
+      end
+    end)
+  end
+
+  # Extract key = {value} or key = "value" pairs from an entry block.
+  # Handles simple single-depth braces; good enough for common fields.
+  @field_regex ~r/(\w+)\s*=\s*(?:\{([^{}]*(?:\{[^{}]*\}[^{}]*)*)\}|"([^"]*)")/
+
+  defp parse_fields(chunk) do
+    @field_regex
+    |> Regex.scan(chunk)
+    |> Enum.reduce(%{}, fn match, acc ->
+      field_name = Enum.at(match, 1) |> String.downcase()
+      # Value is in capture group 2 (braces) or 3 (quotes)
+      value =
+        case {Enum.at(match, 2, ""), Enum.at(match, 3, "")} do
+          {"", q} -> q
+          {b, _} -> b
+        end
+
+      Map.put(acc, field_name, String.trim(value))
+    end)
+  end
+
+  # ------------------------------------------------------------------
+  # Label & URL building
+  # ------------------------------------------------------------------
+
+  defp build_label(entry) do
+    author_part =
+      entry
+      |> Map.get("author", "")
+      |> parse_authors()
+      |> format_authors()
+
+    year = Map.get(entry, "year", Map.get(entry, "date", ""))
+    year = extract_year(year)
+
+    if year && author_part != "", do: "#{author_part}, #{year}", else: author_part
+  end
+
+  defp parse_authors(""), do: []
+
+  defp parse_authors(author_str) do
+    author_str
+    |> String.split(" and ", trim: true)
+    |> Enum.map(&extract_last_name/1)
+    |> Enum.reject(&(&1 == ""))
+  end
+
+  # Handles "Last, First" and "First Last" formats
+  defp extract_last_name(name) do
+    name = String.trim(name)
+
+    cond do
+      String.contains?(name, ",") ->
+        name |> String.split(",") |> List.first() |> String.trim()
+
+      String.contains?(name, " ") ->
+        name |> String.split(" ") |> List.last() |> String.trim()
+
+      true ->
+        name
+    end
+  end
+
+  defp format_authors([]), do: "Unknown"
+  defp format_authors([single]), do: single
+  defp format_authors([first | rest]), do: "#{first} & #{List.last(rest)}"
+
+  defp extract_year(""), do: nil
+
+  defp extract_year(str) do
+    case Regex.run(~r/\b(\d{4})\b/, str) do
+      [_, year] -> year
+      _ -> nil
+    end
+  end
+
+  defp build_url(entry) do
+    cond do
+      doi = Map.get(entry, "doi", "") |> non_empty() ->
+        "https://doi.org/#{doi}"
+
+      url = Map.get(entry, "url", "") |> non_empty() ->
+        url
+
+      true ->
+        nil
+    end
+  end
+
+  defp non_empty(""), do: nil
+  defp non_empty(v), do: v
+end
--- a/org-garden/lib/org_garden/resolvers/doi.ex
+++ b/org-garden/lib/org_garden/resolvers/doi.ex
@@ -0,0 +1,18 @@
+defmodule OrgGarden.Resolvers.DOI do
+  @moduledoc """
+  Last-resort citation resolver — always succeeds.
+
+  If the citation key looks like a DOI (starts with "10."), returns a
+  `https://doi.org/...` link. Otherwise returns the key itself as a
+  plain label with no URL.
+  """
+
+  @spec resolve(String.t()) :: {:ok, map()}
+  def resolve(key) do
+    if String.starts_with?(key, "10.") do
+      {:ok, %{label: key, url: "https://doi.org/#{key}"}}
+    else
+      {:ok, %{label: key, url: nil}}
+    end
+  end
+end
--- a/org-garden/lib/org_garden/resolvers/zotero.ex
+++ b/org-garden/lib/org_garden/resolvers/zotero.ex
@@ -0,0 +1,182 @@
+defmodule OrgGarden.Resolvers.Zotero do
+  @moduledoc """
+  Resolves citation keys via Zotero Better BibTeX's JSON-RPC API.
+
+  Requires Zotero to be running with the Better BibTeX plugin installed.
+  Default endpoint: http://localhost:23119/better-bibtex/json-rpc
+
+  Resolution strategy:
+    1. Search by citation key via `item.search`
+    2. If found, try to get a PDF attachment link (zotero://open-pdf/...)
+    3. Fall back to zotero://select/items/@key
+
+  Returns `{:ok, %{label: "Author, Year", url: "zotero://..."}}` or `:error`.
+  """
+
+  require Logger
+
+  @rpc_path "/better-bibtex/json-rpc"
+
+  @doc """
+  Attempt to resolve `key` against a running Zotero instance.
+  `base_url` defaults to `http://localhost:23119`.
+  """
+  @spec resolve(String.t(), String.t()) :: {:ok, map()} | :error
+  def resolve(key, base_url \\ "http://localhost:23119") do
+    url = base_url <> @rpc_path
+
+    payload =
+      Jason.encode!(%{
+        jsonrpc: "2.0",
+        method: "item.search",
+        params: [
+          [["citationKey", "is", key]]
+        ],
+        id: 1
+      })
+
+    case Req.post(url,
+           body: payload,
+           headers: [{"content-type", "application/json"}],
+           receive_timeout: 5_000,
+           finch: OrgGarden.Finch
+         ) do
+      {:ok, %{status: 200, body: body}} ->
+        parse_response(body, key, base_url)
+
+      {:ok, %{status: status}} ->
+        Logger.debug("Zotero: unexpected HTTP #{status} for key #{key}")
+        :error
+
+      {:error, reason} ->
+        Logger.debug("Zotero: connection failed for key #{key}: #{inspect(reason)}")
+        :error
+
+      other ->
+        Logger.debug("Zotero: unexpected result for key #{key}: #{inspect(other)}")
+        :error
+    end
+  rescue
+    e ->
+      Logger.debug("Zotero: exception resolving key #{key}: #{inspect(e)}")
+      :error
+  end
+
+  # ------------------------------------------------------------------
+  # Private helpers
+  # ------------------------------------------------------------------
+
+  defp parse_response(%{"result" => [item | _]}, key, base_url) do
+    label = build_label(item)
+    url = resolve_url(item, key, base_url)
+    {:ok, %{label: label, url: url}}
+  end
+
+  defp parse_response(%{"result" => []}, key, _base_url) do
+    Logger.debug("Zotero: no item found for key #{key}")
+    :error
+  end
+
+  defp parse_response(%{"error" => err}, key, _base_url) do
+    Logger.debug("Zotero: RPC error for key #{key}: #{inspect(err)}")
+    :error
+  end
+
+  defp parse_response(body, key, _base_url) do
+    Logger.debug("Zotero: unexpected response shape for key #{key}: #{inspect(body)}")
+    :error
+  end
+
+  defp fetch_pdf_url(key, base_url) do
+    payload =
+      Jason.encode!(%{
+        jsonrpc: "2.0",
+        method: "item.attachments",
+        params: [key],
+        id: 2
+      })
+
+    case Req.post(base_url <> @rpc_path,
+           body: payload,
+           headers: [{"content-type", "application/json"}],
+           receive_timeout: 5_000,
+           finch: OrgGarden.Finch
+         ) do
+      {:ok, %{status: 200, body: %{"result" => attachments}}} when is_list(attachments) ->
+        attachments
+        |> Enum.find_value(fn att ->
+          open = Map.get(att, "open", "")
+          path = Map.get(att, "path", "")
+          if String.ends_with?(path, ".pdf"), do: open, else: nil
+        end)
+
+      _ ->
+        nil
+    end
+  rescue
+    _ -> nil
+  end
+
+  # CSL-JSON format: authors are under "author" with "family"/"given" keys.
+  # Year is under "issued" -> "date-parts" -> [[year, month, day]].
+  defp build_label(item) do
+    authors = Map.get(item, "author", [])
+    year = extract_year(item)
+
+    author_part =
+      case authors do
+        [] ->
+          "Unknown"
+
+        [single] ->
+          Map.get(single, "family", Map.get(single, "literal", "Unknown"))
+
+        [first | rest] ->
+          first_name = Map.get(first, "family", Map.get(first, "literal", "Unknown"))
+          last_name =
+            rest
+            |> List.last()
+            |> then(&Map.get(&1, "family", Map.get(&1, "literal", "Unknown")))
+
+          "#{first_name} & #{last_name}"
+      end
+
+    if year, do: "#{author_part}, #{year}", else: author_part
+  end
+
+  # "issued": {"date-parts": [["2021", 2, 3]]}
+  defp extract_year(item) do
+    case get_in(item, ["issued", "date-parts"]) do
+      [[year | _] | _] -> to_string(year)
+      _ -> nil
+    end
+  end
+
+  defp resolve_url(item, key, base_url) do
+    # Prefer zotero://open-pdf/... for items with a PDF attachment.
+    # Fall back to zotero://select/library/items/KEY to open the item in Zotero.
+    # The "id" field is a URI like "http://zotero.org/users/123/items/ABCD1234".
+    pdf_url = fetch_pdf_url(key, base_url)
+
+    if pdf_url do
+      pdf_url
+    else
+      item_key =
+        item
+        |> Map.get("id", "")
+        |> String.split("/")
+        |> List.last()
+        |> non_empty()
+
+      if item_key do
+        "zotero://select/library/items/#{item_key}"
+      else
+        "zotero://select/items/@#{key}"
+      end
+    end
+  end
+
+  defp non_empty(nil), do: nil
+  defp non_empty(""), do: nil
+  defp non_empty(v), do: v
+end
--- a/org-garden/lib/org_garden/supervisor.ex
+++ b/org-garden/lib/org_garden/supervisor.ex
@@ -0,0 +1,40 @@
+defmodule OrgGarden.Supervisor do
+  @moduledoc """
+  Supervises development server components.
+
+  Strategy: :one_for_all
+  If either child fails, restart both to ensure consistent state.
+
+  Children:
+    1. OrgGarden.Watcher - watches .org files for changes
+    2. OrgGarden.Quartz  - runs Quartz Node.js server
+
+  ## Usage
+
+      OrgGarden.Supervisor.start_link(
+        notes_dir: "/path/to/notes",
+        output_dir: "/path/to/output",
+        content_dir: "/path/to/output/content",
+        pipeline_opts: %{zotero_url: "...", ...},
+        transforms: [OrgGarden.Transforms.Citations],
+        port: 8080,
+        ws_port: 3001
+      )
+  """
+  use Supervisor
+
+  def start_link(opts) do
+    Supervisor.start_link(__MODULE__, opts, name: __MODULE__)
+  end
+
+  @impl true
+  def init(opts) do
+    children = [
+      {OrgGarden.Watcher,
+       Keyword.take(opts, [:notes_dir, :output_dir, :content_dir, :pipeline_opts, :transforms])},
+      {OrgGarden.Quartz, Keyword.take(opts, [:content_dir, :port, :ws_port])}
+    ]
+
+    Supervisor.init(children, strategy: :one_for_all)
+  end
+end
--- a/org-garden/lib/org_garden/transform.ex
+++ b/org-garden/lib/org_garden/transform.ex
@@ -0,0 +1,48 @@
+defmodule OrgGarden.Transform do
+  @moduledoc """
+  Behaviour that all markdown transform modules must implement.
+
+  ## Callbacks
+
+  - `init/1`      — called once before processing; returns transform-specific state.
+                    Default implementation returns the opts map unchanged.
+  - `apply/3`     — called per .md file; returns the (possibly modified) content.
+  - `teardown/1`  — optional cleanup after all files are processed.
+
+  ## Example
+
+      defmodule MyTransform do
+        @behaviour OrgGarden.Transform
+
+        @impl true
+        def init(opts), do: %{some_state: opts[:value]}
+
+        @impl true
+        def apply(content, state, _opts) do
+          String.replace(content, "foo", state.some_state)
+        end
+      end
+  """
+
+  @doc "One-time initialisation. Returns opaque state passed to apply/3."
+  @callback init(opts :: map()) :: term()
+
+  @doc "Transform file content. Returns the (possibly modified) content string."
+  @callback apply(content :: String.t(), state :: term(), opts :: map()) :: String.t()
+
+  @doc "Optional cleanup after all files are processed."
+  @callback teardown(state :: term()) :: :ok
+
+  @optional_callbacks teardown: 1
+
+  defmacro __using__(_) do
+    quote do
+      @behaviour OrgGarden.Transform
+
+      @impl OrgGarden.Transform
+      def init(opts), do: opts
+
+      defoverridable init: 1
+    end
+  end
+end
--- a/org-garden/lib/org_garden/transforms/citations.ex
+++ b/org-garden/lib/org_garden/transforms/citations.ex
@@ -0,0 +1,231 @@
+defmodule OrgGarden.Transforms.Citations do
+  @moduledoc """
+  Markdown transform: resolves org-citar citation keys to hyperlinks.
+
+  ## Recognised citation syntax (as output by ox-hugo from org-citar)
+
+      [cite:@key]           → org-cite / citar standard (most common)
+      [cite:@key1;@key2]    → multiple citations
+      cite:key              → older roam-style bare cite syntax
+
+  ## Resolution chain (in order)
+
+    1. Zotero (live instance via Better BibTeX JSON-RPC)   — preferred
+    2. BibTeX file (BIBTEX_FILE env var)                   — fallback
+    3. DOI / bare key                                       — always succeeds
+
+  ## Modes (opts.citation_mode)
+
+    :silent  — silently use DOI/bare-key fallback when Zotero+BibTeX fail
+    :warn    — (default) emit a Logger.warning for unresolved keys
+    :strict  — raise on unresolved keys (aborts pipeline)
+
+  ## Format
+
+  Resolved citations are rendered as:
+
+      [Label](url)       when a URL is available
+      [Label]            when no URL could be determined (bare key fallback)
+
+  Multiple semicolon-separated keys become space-separated links:
+
+      [cite:@a;@b]  →  [Author A, 2020](url_a) [Author B, 2019](url_b)
+
+  ## init/1 callback
+
+  Loads the BibTeX file (if configured) once before processing begins,
+  and probes Zotero availability, emitting warnings as appropriate.
+  """
+
+  @behaviour OrgGarden.Transform
+
+  require Logger
+
+  alias OrgGarden.Resolvers.Zotero
+  alias OrgGarden.Resolvers.BibTeX
+  alias OrgGarden.Resolvers.DOI
+
+  # Match [cite:@key] and [cite:@key1;@key2;...] (org-cite / citar style)
+  @cite_bracket_regex ~r/\[cite:(@[^\]]+)\]/
+
+  # Match bare cite:key or cite:@key (older roam style, no brackets, optional @ prefix)
+  @cite_bare_regex ~r/(?<![(\[])cite:@?([a-zA-Z0-9_:-]+)/
+
+  # ------------------------------------------------------------------
+  # OrgGarden callbacks
+  # ------------------------------------------------------------------
+
+  @doc """
+  Called once before processing any files. Loads BibTeX, probes Zotero.
+  Returns a state map passed to every `apply/3` call.
+  """
+  def init(opts) do
+    bibtex_entries = load_bibtex(opts)
+    zotero_available = probe_zotero(opts)
+
+    if not zotero_available and bibtex_entries == %{} do
+      Logger.warning(
+        "Citations: neither Zotero nor a BibTeX file is available. " <>
+          "All citations will fall back to bare-key rendering. " <>
+          "Set BIBTEX_FILE env var or start Zotero with Better BibTeX to resolve citations."
+      )
+    end
+
+    %{
+      bibtex_entries: bibtex_entries,
+      zotero_available: zotero_available,
+      zotero_url: Map.get(opts, :zotero_url, "http://localhost:23119"),
+      citation_mode: Map.get(opts, :citation_mode, :warn)
+    }
+  end
+
+  @doc """
+  Apply citation resolution to a single markdown file's content.
+  """
+  def apply(content, state, _opts) do
+    content
+    |> resolve_bracket_citations(state)
+    |> resolve_bare_citations(state)
+  end
+
+  # ------------------------------------------------------------------
+  # Resolution passes
+  # ------------------------------------------------------------------
+
+  defp resolve_bracket_citations(content, state) do
+    Regex.replace(@cite_bracket_regex, content, fn _full, keys_str ->
+      keys_str
+      |> String.split(";")
+      |> Enum.map(&String.trim/1)
+      |> Enum.map(fn "@" <> key -> key end)
+      |> Enum.map(&resolve_key(&1, state))
+      |> Enum.join(" ")
+    end)
+  end
+
+  defp resolve_bare_citations(content, state) do
+    Regex.replace(@cite_bare_regex, content, fn _full, key ->
+      resolve_key(key, state)
+    end)
+  end
+
+  # ------------------------------------------------------------------
+  # Single-key resolution chain
+  # ------------------------------------------------------------------
+
+  defp resolve_key(key, state) do
+    info =
+      with :error <- try_zotero(key, state),
+           :error <- try_bibtex(key, state) do
+        handle_unresolved(key, state)
+      else
+        {:ok, citation_info} -> citation_info
+      end
+
+    format_result(info)
+  end
+
+  defp try_zotero(_key, %{zotero_available: false}), do: :error
+
+  defp try_zotero(key, %{zotero_url: url}) do
+    Zotero.resolve(key, url)
+  end
+
+  defp try_bibtex(_key, %{bibtex_entries: entries}) when map_size(entries) == 0, do: :error
+
+  defp try_bibtex(key, %{bibtex_entries: entries}) do
+    BibTeX.resolve(key, entries)
+  end
+
+  defp handle_unresolved(key, %{citation_mode: mode}) do
+    case mode do
+      :strict ->
+        raise "Citations: could not resolve citation key '#{key}' and mode is :strict"
+
+      :warn ->
+        Logger.warning("Citations: unresolved citation key '#{key}' — using bare-key fallback")
+        {:ok, result} = DOI.resolve(key)
+        result
+
+      :silent ->
+        {:ok, result} = DOI.resolve(key)
+        result
+    end
+  end
+
+  defp format_result(%{label: label, url: nil}), do: "[#{label}]"
+  defp format_result(%{label: label, url: url}), do: "[#{label}](#{url})"
+
+  # ------------------------------------------------------------------
+  # Init helpers
+  # ------------------------------------------------------------------
+
+  defp load_bibtex(opts) do
+    path = Map.get(opts, :bibtex_file) || System.get_env("BIBTEX_FILE")
+
+    cond do
+      is_nil(path) ->
+        Logger.debug("Citations: BIBTEX_FILE not set — BibTeX resolver disabled")
+        %{}
+
+      not File.exists?(path) ->
+        Logger.warning("Citations: BIBTEX_FILE=#{path} does not exist — BibTeX resolver disabled")
+        %{}
+
+      true ->
+        case BibTeX.load(path) do
+          {:ok, entries} -> entries
+          {:error, reason} ->
+            Logger.warning("Citations: failed to load BibTeX file #{path}: #{inspect(reason)}")
+            %{}
+        end
+    end
+  end
+
+  defp probe_zotero(opts) do
+    url = Map.get(opts, :zotero_url, "http://localhost:23119")
+
+    # Use a no-op JSON-RPC call to probe availability.
+    # /better-bibtex/cayw is intentionally avoided — it blocks waiting for
+    # user interaction and never returns without a pick.
+    payload =
+      Jason.encode!(%{
+        jsonrpc: "2.0",
+        method: "item.search",
+        params: [[[]]],
+        id: 0
+      })
+
+    result =
+      try do
+        Req.post(url <> "/better-bibtex/json-rpc",
+          body: payload,
+          headers: [{"content-type", "application/json"}],
+          receive_timeout: 3_000,
+          finch: OrgGarden.Finch
+        )
+      rescue
+        e -> {:error, e}
+      end
+
+    case result do
+      {:ok, %{status: 200}} ->
+        Logger.info("Citations: Zotero Better BibTeX is available at #{url}")
+        true
+
+      {:ok, %{status: status}} ->
+        Logger.warning(
+          "Citations: Zotero responded HTTP #{status} at #{url} — " <>
+            "is Better BibTeX installed?"
+        )
+        false
+
+      _ ->
+        Logger.warning(
+          "Citations: Zotero not reachable at #{url} — " <>
+            "start Zotero with Better BibTeX or set BIBTEX_FILE as fallback"
+        )
+        false
+    end
+  end
+end
--- a/org-garden/lib/org_garden/watcher.ex
+++ b/org-garden/lib/org_garden/watcher.ex
@@ -0,0 +1,236 @@
+defmodule OrgGarden.Watcher do
+  @moduledoc """
+  File-watching GenServer that detects `.org` file changes and triggers
+  incremental export + transform for only the affected files.
+
+  Uses the `file_system` package (inotify on Linux, fsevents on macOS)
+  to watch the notes directory. Events are debounced per-file (500ms)
+  to coalesce rapid writes (e.g., Emacs auto-save).
+
+  ## Lifecycle
+
+  Started dynamically by `OrgGarden.CLI` after the initial batch export.
+  Transforms are initialized once at startup and reused across all
+  incremental rebuilds to avoid repeated Zotero probes and BibTeX loads.
+
+  ## Usage
+
+      OrgGarden.Watcher.start_link(
+        notes_dir: "/path/to/notes",
+        output_dir: "/path/to/output",
+        content_dir: "/path/to/output/content",
+        pipeline_opts: %{zotero_url: "...", ...},
+        transforms: [OrgGarden.Transforms.Citations]
+      )
+  """
+
+  use GenServer
+
+  require Logger
+
+  @debounce_ms 500
+
+  # -------------------------------------------------------------------
+  # Client API
+  # -------------------------------------------------------------------
+
+  @doc """
+  Start the watcher as a linked process.
+
+  ## Options
+
+    * `:notes_dir`      — directory to watch for `.org` changes (required)
+    * `:output_dir`     — ox-hugo base dir (required)
+    * `:content_dir`    — directory where `.md` files are written (required)
+    * `:pipeline_opts`  — opts map passed to transforms (required)
+    * `:transforms`     — list of transform modules (default: `[OrgGarden.Transforms.Citations]`)
+  """
+  def start_link(opts) do
+    GenServer.start_link(__MODULE__, opts, name: __MODULE__)
+  end
+
+  # -------------------------------------------------------------------
+  # GenServer callbacks
+  # -------------------------------------------------------------------
+
+  @impl true
+  def init(opts) do
+    notes_dir = Keyword.fetch!(opts, :notes_dir)
+    output_dir = Keyword.fetch!(opts, :output_dir)
+    content_dir = Keyword.fetch!(opts, :content_dir)
+    pipeline_opts = Keyword.fetch!(opts, :pipeline_opts)
+    transforms = Keyword.get(opts, :transforms, [OrgGarden.Transforms.Citations])
+
+    # Initialize transforms once — reused for all incremental rebuilds
+    initialized_transforms = OrgGarden.init_transforms(transforms, pipeline_opts)
+
+    # Start the file system watcher
+    {:ok, watcher_pid} = FileSystem.start_link(dirs: [notes_dir], recursive: true)
+    FileSystem.subscribe(watcher_pid)
+
+    Logger.info("Watcher: monitoring #{notes_dir} for .org changes")
+
+    {:ok,
+     %{
+       notes_dir: notes_dir,
+       output_dir: output_dir,
+       content_dir: content_dir,
+       pipeline_opts: pipeline_opts,
+       watcher_pid: watcher_pid,
+       initialized_transforms: initialized_transforms,
+       pending: %{}
+     }}
+  end
+
+  @impl true
+  def handle_info({:file_event, _pid, {path, events}}, state) do
+    path = to_string(path)
+
+    if org_file?(path) and not temporary_file?(path) do
+      event_type = classify_events(events)
+      Logger.debug("Watcher: #{event_type} event for #{path}")
+      {:noreply, schedule_debounce(path, event_type, state)}
+    else
+      {:noreply, state}
+    end
+  end
+
+  @impl true
+  def handle_info({:file_event, _pid, :stop}, state) do
+    Logger.warning("Watcher: file system monitor stopped unexpectedly")
+    {:stop, :watcher_stopped, state}
+  end
+
+  @impl true
+  def handle_info({:debounced, path, event_type}, state) do
+    state = %{state | pending: Map.delete(state.pending, path)}
+
+    case event_type do
+      :deleted ->
+        handle_delete(path, state)
+
+      _created_or_modified ->
+        handle_change(path, state)
+    end
+
+    {:noreply, state}
+  end
+
+  @impl true
+  def terminate(_reason, state) do
+    OrgGarden.teardown_transforms(state.initialized_transforms)
+    :ok
+  end
+
+  # -------------------------------------------------------------------
+  # Event handling
+  # -------------------------------------------------------------------
+
+  defp handle_change(orgfile, state) do
+    %{
+      notes_dir: notes_dir,
+      output_dir: output_dir,
+      content_dir: content_dir,
+      pipeline_opts: pipeline_opts,
+      initialized_transforms: initialized_transforms
+    } = state
+
+    md_path = OrgGarden.Export.expected_md_path(orgfile, notes_dir, content_dir)
+    IO.puts("==> Changed: #{Path.relative_to(orgfile, notes_dir)}")
+
+    case OrgGarden.Export.export_file(orgfile, notes_dir, output_dir) do
+      {:ok, _} ->
+        IO.puts("  exported: #{Path.relative_to(md_path, content_dir)}")
+
+        {:ok, stats} = OrgGarden.run_on_files_with([md_path], initialized_transforms, pipeline_opts)
+
+        Enum.each(stats, fn {mod, count} ->
+          if count > 0, do: IO.puts("  #{inspect(mod)}: #{count} file(s) modified")
+        end)
+
+        regenerate_index(content_dir)
+        IO.puts("==> Done")
+
+      {:error, reason} ->
+        Logger.error("Watcher: export failed for #{orgfile}: #{inspect(reason)}")
+    end
+  end
+
+  defp handle_delete(orgfile, state) do
+    %{notes_dir: notes_dir, content_dir: content_dir} = state
+
+    md_path = OrgGarden.Export.expected_md_path(orgfile, notes_dir, content_dir)
+    IO.puts("==> Deleted: #{Path.relative_to(orgfile, notes_dir)}")
+
+    if File.exists?(md_path) do
+      File.rm!(md_path)
+      IO.puts("  removed: #{Path.relative_to(md_path, content_dir)}")
+
+      # Clean up empty parent directories left behind
+      cleanup_empty_dirs(Path.dirname(md_path), content_dir)
+    end
+
+    regenerate_index(content_dir)
+    IO.puts("==> Done")
+  end
+
+  # -------------------------------------------------------------------
+  # Index generation
+  # -------------------------------------------------------------------
+
+  defp regenerate_index(content_dir) do
+    OrgGarden.Index.regenerate(content_dir)
+  end
+
+  # -------------------------------------------------------------------
+  # Helpers
+  # -------------------------------------------------------------------
+
+  defp schedule_debounce(path, event_type, state) do
+    # Cancel any existing timer for this path
+    case Map.get(state.pending, path) do
+      nil -> :ok
+      old_ref -> Process.cancel_timer(old_ref)
+    end
+
+    ref = Process.send_after(self(), {:debounced, path, event_type}, @debounce_ms)
+    %{state | pending: Map.put(state.pending, path, ref)}
+  end
+
+  defp org_file?(path), do: String.ends_with?(path, ".org")
+
+  defp temporary_file?(path) do
+    basename = Path.basename(path)
+    # Emacs creates temp files like .#file.org and #file.org#
+    String.starts_with?(basename, ".#") or
+      (String.starts_with?(basename, "#") and String.ends_with?(basename, "#"))
+  end
+
+  defp classify_events(events) do
+    cond do
+      :removed in events or :deleted in events -> :deleted
+      :created in events -> :created
+      :modified in events or :changed in events -> :modified
+      # renamed can mean created or deleted depending on context;
+      # if the file exists it was renamed into the watched dir
+      :renamed in events -> :modified
+      true -> :modified
+    end
+  end
+
+  defp cleanup_empty_dirs(dir, stop_at) do
+    dir = Path.expand(dir)
+    stop_at = Path.expand(stop_at)
+
+    if dir != stop_at and File.dir?(dir) do
+      case File.ls!(dir) do
+        [] ->
+          File.rmdir!(dir)
+          cleanup_empty_dirs(Path.dirname(dir), stop_at)
+
+        _ ->
+          :ok
+      end
+    end
+  end
+end
--- a/org-garden/mix.exs
+++ b/org-garden/mix.exs
@@ -0,0 +1,34 @@
+defmodule OrgGarden.MixProject do
+  use Mix.Project
+
+  def project do
+    [
+      app: :org_garden,
+      version: "0.1.0",
+      elixir: "~> 1.17",
+      start_permanent: Mix.env() == :prod,
+      deps: deps(),
+      escript: escript()
+    ]
+  end
+
+  def application do
+    [
+      extra_applications: [:logger],
+      mod: {OrgGarden.Application, []}
+    ]
+  end
+
+  defp escript do
+    [main_module: OrgGarden.CLI]
+  end
+
+  defp deps do
+    [
+      {:finch, "~> 0.19"},
+      {:req, "~> 0.5"},
+      {:jason, "~> 1.4"},
+      {:file_system, "~> 1.0"}
+    ]
+  end
+end
--- a/org-garden/mix.lock
+++ b/org-garden/mix.lock
@@ -0,0 +1,12 @@
+%{
+  "file_system": {:hex, :file_system, "1.1.1", "31864f4685b0148f25bd3fbef2b1228457c0c89024ad67f7a81a3ffbc0bbad3a", [:mix], [], "hexpm", "7a15ff97dfe526aeefb090a7a9d3d03aa907e100e262a0f8f7746b78f8f87a5d"},
+  "finch": {:hex, :finch, "0.21.0", "b1c3b2d48af02d0c66d2a9ebfb5622be5c5ecd62937cf79a88a7f98d48a8290c", [:mix], [{:mime, "~> 1.0 or ~> 2.0", [hex: :mime, repo: "hexpm", optional: false]}, {:mint, "~> 1.6.2 or ~> 1.7", [hex: :mint, repo: "hexpm", optional: false]}, {:nimble_options, "~> 0.4 or ~> 1.0", [hex: :nimble_options, repo: "hexpm", optional: false]}, {:nimble_pool, "~> 1.1", [hex: :nimble_pool, repo: "hexpm", optional: false]}, {:telemetry, "~> 0.4 or ~> 1.0", [hex: :telemetry, repo: "hexpm", optional: false]}], "hexpm", "87dc6e169794cb2570f75841a19da99cfde834249568f2a5b121b809588a4377"},
+  "hpax": {:hex, :hpax, "1.0.3", "ed67ef51ad4df91e75cc6a1494f851850c0bd98ebc0be6e81b026e765ee535aa", [:mix], [], "hexpm", "8eab6e1cfa8d5918c2ce4ba43588e894af35dbd8e91e6e55c817bca5847df34a"},
+  "jason": {:hex, :jason, "1.4.4", "b9226785a9aa77b6857ca22832cffa5d5011a667207eb2a0ad56adb5db443b8a", [:mix], [{:decimal, "~> 1.0 or ~> 2.0", [hex: :decimal, repo: "hexpm", optional: true]}], "hexpm", "c5eb0cab91f094599f94d55bc63409236a8ec69a21a67814529e8d5f6cc90b3b"},
+  "mime": {:hex, :mime, "2.0.7", "b8d739037be7cd402aee1ba0306edfdef982687ee7e9859bee6198c1e7e2f128", [:mix], [], "hexpm", "6171188e399ee16023ffc5b76ce445eb6d9672e2e241d2df6050f3c771e80ccd"},
+  "mint": {:hex, :mint, "1.7.1", "113fdb2b2f3b59e47c7955971854641c61f378549d73e829e1768de90fc1abf1", [:mix], [{:castore, "~> 0.1.0 or ~> 1.0", [hex: :castore, repo: "hexpm", optional: true]}, {:hpax, "~> 0.1.1 or ~> 0.2.0 or ~> 1.0", [hex: :hpax, repo: "hexpm", optional: false]}], "hexpm", "fceba0a4d0f24301ddee3024ae116df1c3f4bb7a563a731f45fdfeb9d39a231b"},
+  "nimble_options": {:hex, :nimble_options, "1.1.1", "e3a492d54d85fc3fd7c5baf411d9d2852922f66e69476317787a7b2bb000a61b", [:mix], [], "hexpm", "821b2470ca9442c4b6984882fe9bb0389371b8ddec4d45a9504f00a66f650b44"},
+  "nimble_pool": {:hex, :nimble_pool, "1.1.0", "bf9c29fbdcba3564a8b800d1eeb5a3c58f36e1e11d7b7fb2e084a643f645f06b", [:mix], [], "hexpm", "af2e4e6b34197db81f7aad230c1118eac993acc0dae6bc83bac0126d4ae0813a"},
+  "req": {:hex, :req, "0.5.17", "0096ddd5b0ed6f576a03dde4b158a0c727215b15d2795e59e0916c6971066ede", [:mix], [{:brotli, "~> 0.3.1", [hex: :brotli, repo: "hexpm", optional: true]}, {:ezstd, "~> 1.0", [hex: :ezstd, repo: "hexpm", optional: true]}, {:finch, "~> 0.17", [hex: :finch, repo: "hexpm", optional: false]}, {:jason, "~> 1.0", [hex: :jason, repo: "hexpm", optional: false]}, {:mime, "~> 2.0.6 or ~> 2.1", [hex: :mime, repo: "hexpm", optional: false]}, {:nimble_csv, "~> 1.0", [hex: :nimble_csv, repo: "hexpm", optional: true]}, {:plug, "~> 1.0", [hex: :plug, repo: "hexpm", optional: true]}], "hexpm", "0b8bc6ffdfebbc07968e59d3ff96d52f2202d0536f10fef4dc11dc02a2a43e39"},
+  "telemetry": {:hex, :telemetry, "1.3.0", "fedebbae410d715cf8e7062c96a1ef32ec22e764197f70cda73d82778d61e7a2", [:rebar3], [], "hexpm", "7015fc8919dbe63764f4b4b87a95b7c0996bd539e0d499be6ec9d7f3875b79e6"},
+}
--- a/org-garden/patches/01-glob-gitignore.patch
+++ b/org-garden/patches/01-glob-gitignore.patch
@@ -0,0 +1,19 @@
+diff --git a/quartz/util/glob.ts b/quartz/util/glob.ts
+index 7a71160..91fbaa7 100644
+--- a/quartz/util/glob.ts
+++ b/quartz/util/glob.ts
+@@ -10,12 +10,13 @@ export async function glob(
+   pattern: string,
+   cwd: string,
+   ignorePatterns: string[],
+  respectGitignore: boolean = true,
+ ): Promise<FilePath[]> {
+   const fps = (
+     await globby(pattern, {
+       cwd,
+       ignore: ignorePatterns,
+-      gitignore: true,
+      gitignore: respectGitignore,
+     })
+   ).map(toPosixPath)
+   return fps as FilePath[]
--- a/org-garden/patches/02-build-gitignore.patch
+++ b/org-garden/patches/02-build-gitignore.patch
@@ -0,0 +1,13 @@
+diff --git a/quartz/build.ts b/quartz/build.ts
+index b98f4a8..3166a06 100644
+--- a/quartz/build.ts
+++ b/quartz/build.ts
+@@ -71,7 +71,7 @@ async function buildQuartz(argv: Argv, mut: Mutex, clientRefresh: () => void) {
+   console.log(`Cleaned output directory \`${output}\` in ${perf.timeSince("clean")}`)
+ 
+   perf.addEvent("glob")
+-  const allFiles = await glob("**/*.*", argv.directory, cfg.configuration.ignorePatterns)
+  const allFiles = await glob("**/*.*", argv.directory, cfg.configuration.ignorePatterns, false)
+   const markdownPaths = allFiles.filter((fp) => fp.endsWith(".md")).sort()
+   console.log(
+     `Found ${markdownPaths.length} input files from \`${argv.directory}\` in ${perf.timeSince("glob")}`,
--- a/org-garden/patches/03-static-hugo.patch
+++ b/org-garden/patches/03-static-hugo.patch
@@ -0,0 +1,34 @@
+diff --git a/quartz/plugins/emitters/static.ts b/quartz/plugins/emitters/static.ts
+index 0b45290..8b34049 100644
+--- a/quartz/plugins/emitters/static.ts
+++ b/quartz/plugins/emitters/static.ts
+@@ -7,6 +7,7 @@ import { dirname } from "path"
+ export const Static: QuartzEmitterPlugin = () => ({
+   name: "Static",
+   async *emit({ argv, cfg }) {
+    // Copy Quartz's own internal static assets (quartz/static/) → output/static/
+     const staticPath = joinSegments(QUARTZ, "static")
+     const fps = await glob("**", staticPath, cfg.configuration.ignorePatterns)
+     const outputStaticPath = joinSegments(argv.output, "static")
+@@ -18,6 +19,21 @@ export const Static: QuartzEmitterPlugin = () => ({
+       await fs.promises.copyFile(src, dest)
+       yield dest
+     }
+
+    // Copy user-facing static assets (static/) → output/ preserving paths.
+    // This mirrors Hugo's convention: static/ox-hugo/foo.png is served at /ox-hugo/foo.png,
+    // which matches the src="/ox-hugo/..." paths that ox-hugo writes into exported markdown.
+    const userStaticPath = "static"
+    if (fs.existsSync(userStaticPath)) {
+      const userFps = await glob("**", userStaticPath, cfg.configuration.ignorePatterns, false)
+      for (const fp of userFps) {
+        const src = joinSegments(userStaticPath, fp) as FilePath
+        const dest = joinSegments(argv.output, fp) as FilePath
+        await fs.promises.mkdir(dirname(dest), { recursive: true })
+        await fs.promises.copyFile(src, dest)
+        yield dest
+      }
+    }
+   },
+   async *partialEmit() {},
+ })
--- a/org-garden/patches/04-oxhugofm-figure.patch
+++ b/org-garden/patches/04-oxhugofm-figure.patch
@@ -0,0 +1,44 @@
+diff --git a/quartz/plugins/transformers/oxhugofm.ts b/quartz/plugins/transformers/oxhugofm.ts
+index 303566e..4fb5e2c 100644
+--- a/quartz/plugins/transformers/oxhugofm.ts
+++ b/quartz/plugins/transformers/oxhugofm.ts
+@@ -27,7 +27,10 @@ const defaultOptions: Options = {
+ const relrefRegex = new RegExp(/\[([^\]]+)\]\(\{\{< relref "([^"]+)" >\}\}\)/, "g")
+ const predefinedHeadingIdRegex = new RegExp(/(.*) {#(?:.*)}/, "g")
+ const hugoShortcodeRegex = new RegExp(/{{(.*)}}/, "g")
+-const figureTagRegex = new RegExp(/< ?figure src="(.*)" ?>/, "g")
+// Matches the full Hugo {{< figure src="..." ... >}} shortcode and captures src.
+// Must run before the generic shortcode stripper to avoid partial-match issues
+// with captions that contain HTML (e.g. <span class="figure-number">).
+const figureShortcodeRegex = new RegExp(/{{<\s*figure\b[^}]*\bsrc="([^"]*)"[^}]*>}}/, "g")
+ // \\\\\( -> matches \\(
+ // (.+?) -> Lazy match for capturing the equation
+ // \\\\\) -> matches \\)
+@@ -70,19 +73,19 @@ export const OxHugoFlavouredMarkdown: QuartzTransformerPlugin<Partial<Options>>
+         })
+       }
+ 
+-      if (opts.removeHugoShortcode) {
+      if (opts.replaceFigureWithMdImg) {
+         src = src.toString()
+-        src = src.replaceAll(hugoShortcodeRegex, (_value, ...capture) => {
+-          const [scContent] = capture
+-          return scContent
+        src = src.replaceAll(figureShortcodeRegex, (_value, ...capture) => {
+          const [imgSrc] = capture
+          return `![](${imgSrc})`
+         })
+       }
+ 
+-      if (opts.replaceFigureWithMdImg) {
+      if (opts.removeHugoShortcode) {
+         src = src.toString()
+-        src = src.replaceAll(figureTagRegex, (_value, ...capture) => {
+-          const [src] = capture
+-          return `![](${src})`
+        src = src.replaceAll(hugoShortcodeRegex, (_value, ...capture) => {
+          const [scContent] = capture
+          return scContent
+         })
+       }
+ 
--- a/org-garden/quartz-config/globals.d.ts
+++ b/org-garden/quartz-config/globals.d.ts
@@ -0,0 +1,17 @@
+export declare global {
+  interface Document {
+    addEventListener<K extends keyof CustomEventMap>(
+      type: K,
+      listener: (this: Document, ev: CustomEventMap[K]) => void,
+    ): void
+    removeEventListener<K extends keyof CustomEventMap>(
+      type: K,
+      listener: (this: Document, ev: CustomEventMap[K]) => void,
+    ): void
+    dispatchEvent<K extends keyof CustomEventMap>(ev: CustomEventMap[K] | UIEvent): void
+  }
+  interface Window {
+    spaNavigate(url: URL, isBack: boolean = false)
+    addCleanup(fn: (...args: any[]) => void)
+  }
+}
--- a/org-garden/quartz-config/index.d.ts
+++ b/org-garden/quartz-config/index.d.ts
@@ -0,0 +1,15 @@
+declare module "*.scss" {
+  const content: string
+  export = content
+}
+
+// dom custom event
+interface CustomEventMap {
+  prenav: CustomEvent<{}>
+  nav: CustomEvent<{ url: FullSlug }>
+  themechange: CustomEvent<{ theme: "light" | "dark" }>
+  readermodechange: CustomEvent<{ mode: "on" | "off" }>
+}
+
+type ContentIndex = Record<FullSlug, ContentDetails>
+declare const fetchData: Promise<ContentIndex>
--- a/org-garden/quartz-config/quartz.config.ts
+++ b/org-garden/quartz-config/quartz.config.ts
@@ -0,0 +1,101 @@
+import { QuartzConfig } from "./quartz/cfg"
+import * as Plugin from "./quartz/plugins"
+
+/**
+ * Quartz 4 Configuration
+ *
+ * See https://quartz.jzhao.xyz/configuration for more information.
+ */
+const config: QuartzConfig = {
+  configuration: {
+    pageTitle: "Quartz 4",
+    pageTitleSuffix: "",
+    enableSPA: true,
+    enablePopovers: true,
+    analytics: {
+      provider: "plausible",
+    },
+    locale: "en-US",
+    baseUrl: "quartz.jzhao.xyz",
+    ignorePatterns: ["private", "templates", ".obsidian"],
+    defaultDateType: "modified",
+    theme: {
+      fontOrigin: "googleFonts",
+      cdnCaching: true,
+      typography: {
+        header: "Schibsted Grotesk",
+        body: "Source Sans Pro",
+        code: "IBM Plex Mono",
+      },
+      colors: {
+        lightMode: {
+          light: "#faf8f8",
+          lightgray: "#e5e5e5",
+          gray: "#b8b8b8",
+          darkgray: "#4e4e4e",
+          dark: "#2b2b2b",
+          secondary: "#284b63",
+          tertiary: "#84a59d",
+          highlight: "rgba(143, 159, 169, 0.15)",
+          textHighlight: "#fff23688",
+        },
+        darkMode: {
+          light: "#161618",
+          lightgray: "#393639",
+          gray: "#646464",
+          darkgray: "#d4d4d4",
+          dark: "#ebebec",
+          secondary: "#7b97aa",
+          tertiary: "#84a59d",
+          highlight: "rgba(143, 159, 169, 0.15)",
+          textHighlight: "#b3aa0288",
+        },
+      },
+    },
+  },
+  plugins: {
+    transformers: [
+      Plugin.FrontMatter({ delimiters: "+++", language: "toml" }),
+      Plugin.CreatedModifiedDate({
+        priority: ["frontmatter", "git", "filesystem"],
+      }),
+      Plugin.SyntaxHighlighting({
+        theme: {
+          light: "github-light",
+          dark: "github-dark",
+        },
+        keepBackground: false,
+      }),
+      // OxHugoFlavouredMarkdown must come before GitHubFlavoredMarkdown.
+      // Note: not compatible with ObsidianFlavoredMarkdown — use one or the other.
+      // If ox-hugo exports TOML frontmatter, change FrontMatter to:
+      //   Plugin.FrontMatter({ delims: "+++", language: "toml" })
+      Plugin.OxHugoFlavouredMarkdown(),
+      Plugin.GitHubFlavoredMarkdown(),
+      Plugin.TableOfContents(),
+      Plugin.CrawlLinks({ markdownLinkResolution: "shortest" }),
+      Plugin.Description(),
+      Plugin.Latex({ renderEngine: "katex" }),
+    ],
+    filters: [Plugin.RemoveDrafts()],
+    emitters: [
+      Plugin.AliasRedirects(),
+      Plugin.ComponentResources(),
+      Plugin.ContentPage(),
+      Plugin.FolderPage(),
+      Plugin.TagPage(),
+      Plugin.ContentIndex({
+        enableSiteMap: true,
+        enableRSS: true,
+      }),
+      Plugin.Assets(),
+      Plugin.Static(),
+      Plugin.Favicon(),
+      Plugin.NotFoundPage(),
+      // Comment out CustomOgImages to speed up build time
+      Plugin.CustomOgImages(),
+    ],
+  },
+}
+
+export default config
--- a/org-garden/quartz-config/quartz.layout.ts
+++ b/org-garden/quartz-config/quartz.layout.ts
@@ -0,0 +1,68 @@
+import { PageLayout, SharedLayout } from "./quartz/cfg"
+import * as Component from "./quartz/components"
+
+// components shared across all pages
+export const sharedPageComponents: SharedLayout = {
+  head: Component.Head(),
+  header: [],
+  afterBody: [],
+  footer: Component.Footer({
+    links: {
+      GitHub: "https://github.com/jackyzha0/quartz",
+      "Discord Community": "https://discord.gg/cRFFHYye7t",
+    },
+  }),
+}
+
+// components for pages that display a single page (e.g. a single note)
+export const defaultContentPageLayout: PageLayout = {
+  beforeBody: [
+    Component.ConditionalRender({
+      component: Component.Breadcrumbs(),
+      condition: (page) => page.fileData.slug !== "index",
+    }),
+    Component.ArticleTitle(),
+    Component.ContentMeta(),
+    Component.TagList(),
+  ],
+  left: [
+    Component.PageTitle(),
+    Component.MobileOnly(Component.Spacer()),
+    Component.Flex({
+      components: [
+        {
+          Component: Component.Search(),
+          grow: true,
+        },
+        { Component: Component.Darkmode() },
+        { Component: Component.ReaderMode() },
+      ],
+    }),
+    Component.Explorer(),
+  ],
+  right: [
+    Component.Graph(),
+    Component.DesktopOnly(Component.TableOfContents()),
+    Component.Backlinks(),
+  ],
+}
+
+// components for pages that display lists of pages  (e.g. tags or folders)
+export const defaultListPageLayout: PageLayout = {
+  beforeBody: [Component.Breadcrumbs(), Component.ArticleTitle(), Component.ContentMeta()],
+  left: [
+    Component.PageTitle(),
+    Component.MobileOnly(Component.Spacer()),
+    Component.Flex({
+      components: [
+        {
+          Component: Component.Search(),
+          grow: true,
+        },
+        { Component: Component.Darkmode() },
+      ],
+    }),
+    Component.Explorer(),
+  ],
+  right: [],
+}
--- a/package.json
+++ b/package.json
@@ -17,7 +17,10 @@
    "check": "tsc --noEmit && npx prettier . --check",
    "format": "npx prettier . --write",
    "test": "tsx --test",
-    "profile": "0x -D prof ./quartz/bootstrap-cli.mjs build --concurrency=1"
+    "profile": "0x -D prof ./quartz/bootstrap-cli.mjs build --concurrency=1",
+    "export": "elixir scripts/export.exs",
+    "build:notes": "elixir scripts/export.exs && npx quartz build",
+    "serve:notes": "elixir scripts/export.exs && npx quartz build --serve"
  },
  "engines": {
    "npm": ">=10.9.2",
--- a/quartz.config.ts
+++ b/quartz.config.ts
@@ -55,7 +55,7 @@ const config: QuartzConfig = {
  },
  plugins: {
    transformers: [
-      Plugin.FrontMatter(),
+      Plugin.FrontMatter({ delimiters: "+++", language: "toml" }),
      Plugin.CreatedModifiedDate({
        priority: ["frontmatter", "git", "filesystem"],
      }),
@@ -66,13 +66,16 @@ const config: QuartzConfig = {
        },
        keepBackground: false,
      }),
-      Plugin.ObsidianFlavoredMarkdown({ enableInHtmlEmbed: false }),
+      // OxHugoFlavouredMarkdown must come before GitHubFlavoredMarkdown.
+      // Note: not compatible with ObsidianFlavoredMarkdown — use one or the other.
+      // If ox-hugo exports TOML frontmatter, change FrontMatter to:
+      //   Plugin.FrontMatter({ delims: "+++", language: "toml" })
+      Plugin.OxHugoFlavouredMarkdown(),
      Plugin.GitHubFlavoredMarkdown(),
      Plugin.TableOfContents(),
      Plugin.CrawlLinks({ markdownLinkResolution: "shortest" }),
      Plugin.Description(),
      Plugin.Latex({ renderEngine: "katex" }),
-      Plugin.ObsidianBases(),
    ],
    filters: [Plugin.RemoveDrafts()],
    emitters: [
@@ -91,7 +94,6 @@ const config: QuartzConfig = {
      Plugin.NotFoundPage(),
      // Comment out CustomOgImages to speed up build time
      Plugin.CustomOgImages(),
-      Plugin.BasePage(),
    ],
  },
 }
--- a/quartz/build.ts
+++ b/quartz/build.ts
@@ -71,8 +71,8 @@ async function buildQuartz(argv: Argv, mut: Mutex, clientRefresh: () => void) {
  console.log(`Cleaned output directory \`${output}\` in ${perf.timeSince("clean")}`)

  perf.addEvent("glob")
-  const allFiles = await glob("**/*.*", argv.directory, cfg.configuration.ignorePatterns)
-  const markdownPaths = allFiles.filter((fp) => fp.endsWith(".md") || fp.endsWith(".base")).sort()
+  const allFiles = await glob("**/*.*", argv.directory, cfg.configuration.ignorePatterns, false)
+  const markdownPaths = allFiles.filter((fp) => fp.endsWith(".md")).sort()
  console.log(
    `Found ${markdownPaths.length} input files from \`${argv.directory}\` in ${perf.timeSince("glob")}`,
  )
--- a/quartz/components/BaseViewSelector.tsx
+++ b/quartz/components/BaseViewSelector.tsx
@@ -1,218 +0,0 @@
-import { JSX } from "preact"
-import { QuartzComponent, QuartzComponentConstructor, QuartzComponentProps } from "./types"
-import { classNames } from "../util/lang"
-import { resolveRelative } from "../util/path"
-// @ts-ignore
-import script from "./scripts/base-view-selector.inline"
-import baseViewSelectorStyle from "./styles/baseViewSelector.scss"
-
-const icons: Record<string, JSX.Element> = {
-  table: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <rect width="18" height="18" x="3" y="3" rx="2" />
-      <path d="M3 9h18" />
-      <path d="M3 15h18" />
-      <path d="M9 3v18" />
-      <path d="M15 3v18" />
-    </svg>
-  ),
-  list: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <line x1="8" x2="21" y1="6" y2="6" />
-      <line x1="8" x2="21" y1="12" y2="12" />
-      <line x1="8" x2="21" y1="18" y2="18" />
-      <line x1="3" x2="3.01" y1="6" y2="6" />
-      <line x1="3" x2="3.01" y1="12" y2="12" />
-      <line x1="3" x2="3.01" y1="18" y2="18" />
-    </svg>
-  ),
-  chevronsUpDown: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <path d="m7 15 5 5 5-5" />
-      <path d="m7 9 5-5 5 5" />
-    </svg>
-  ),
-  chevronRight: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <path d="m9 18 6-6-6-6" />
-    </svg>
-  ),
-  x: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <path d="M18 6 6 18" />
-      <path d="m6 6 12 12" />
-    </svg>
-  ),
-  map: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <path d="M15 6v12a3 3 0 1 0 3-3H6a3 3 0 1 0 3 3V6a3 3 0 1 0-3 3h12a3 3 0 1 0-3-3" />
-    </svg>
-  ),
-  card: (
-    <svg
-      xmlns="http://www.w3.org/2000/svg"
-      width="16"
-      height="16"
-      viewBox="0 0 24 24"
-      fill="none"
-      stroke="currentColor"
-      stroke-width="2"
-      stroke-linecap="round"
-      stroke-linejoin="round"
-    >
-      <rect width="7" height="7" x="3" y="3" rx="1" />
-      <rect width="7" height="7" x="14" y="3" rx="1" />
-      <rect width="7" height="7" x="14" y="14" rx="1" />
-      <rect width="7" height="7" x="3" y="14" rx="1" />
-    </svg>
-  ),
-}
-
-const viewTypeIcons: Record<string, JSX.Element | undefined> = {
-  table: icons.table,
-  list: icons.list,
-  gallery: icons.card,
-  board: icons.table,
-  calendar: icons.table,
-  map: icons.map,
-  cards: icons.card,
-}
-
-const BaseViewSelector: QuartzComponent = ({ fileData, displayClass }: QuartzComponentProps) => {
-  const baseMeta = fileData.basesMetadata
-
-  if (!baseMeta || baseMeta.allViews.length <= 1) {
-    return null
-  }
-
-  const currentViewName = baseMeta.currentView
-  const allViews = baseMeta.allViews
-  const currentIcon =
-    viewTypeIcons[allViews.find((view) => view.name === currentViewName)?.type ?? ""] ?? icons.table
-
-  return (
-    <div class={classNames(displayClass, "bases-toolbar")} data-base-view-selector>
-      <div class="bases-toolbar-item bases-toolbar-views-menu">
-        <span
-          class="text-icon-button"
-          aria-label="Select view"
-          aria-expanded="false"
-          aria-haspopup="true"
-          role="button"
-          tabindex={0}
-        >
-          <span class="text-button-icon">{currentIcon}</span>
-          <span class="text-button-label">{currentViewName.toLowerCase()}</span>
-          <span class="text-button-icon mod-aux">{icons.chevronsUpDown}</span>
-        </span>
-      </div>
-
-      <div class="menu-scroll" data-dropdown>
-        <div class="bases-toolbar-menu-container">
-          <div class="search-input-container">
-            <input type="search" placeholder="Search..." data-search-input />
-            <div class="search-input-clear-button" data-clear-search hidden>
-              {icons.x}
-            </div>
-          </div>
-          <div class="bases-toolbar-items">
-            <div class="suggestion-group" data-group="views" data-view-list>
-              {allViews.map((view) => {
-                const isActive = view.name === currentViewName
-                const icon = viewTypeIcons[view.type] || icons.table
-                const href = resolveRelative(fileData.slug!, view.slug)
-
-                return (
-                  <a
-                    href={href}
-                    data-slug={view.slug}
-                    class={
-                      isActive
-                        ? "suggestion-item bases-toolbar-menu-item mod-active is-selected"
-                        : "suggestion-item bases-toolbar-menu-item"
-                    }
-                    data-view-name={view.name}
-                    data-view-type={view.type}
-                  >
-                    <div class="bases-toolbar-menu-item-info">
-                      <div class="bases-toolbar-menu-item-info-icon">{icon}</div>
-                      <div class="bases-toolbar-menu-item-name">{view.name.toLowerCase()}</div>
-                    </div>
-                    <div class="clickable-icon bases-toolbar-menu-item-icon">
-                      {icons.chevronRight}
-                    </div>
-                  </a>
-                )
-              })}
-            </div>
-          </div>
-        </div>
-      </div>
-    </div>
-  )
-}
-
-BaseViewSelector.css = baseViewSelectorStyle
-BaseViewSelector.afterDOMLoaded = script
-
-export default (() => BaseViewSelector) satisfies QuartzComponentConstructor
--- a/quartz/components/Breadcrumbs.tsx
+++ b/quartz/components/Breadcrumbs.tsx
@@ -51,9 +51,7 @@ export default ((opts?: Partial<BreadcrumbOptions>) => {
    ctx,
  }: QuartzComponentProps) => {
    const trie = (ctx.trie ??= trieFromAllFiles(allFiles))
-    const baseMeta = fileData.basesMetadata
-
-    const slugParts = (baseMeta ? baseMeta.baseSlug : fileData.slug!).split("/")
+    const slugParts = fileData.slug!.split("/")
    const pathNodes = trie.ancestryChain(slugParts)

    if (!pathNodes) {
@@ -66,24 +64,14 @@ export default ((opts?: Partial<BreadcrumbOptions>) => {
        crumb.displayName = options.rootName
      }

+      // For last node (current page), set empty path
      if (idx === pathNodes.length - 1) {
-        if (baseMeta) {
-          crumb.path = resolveRelative(fileData.slug!, simplifySlug(baseMeta.baseSlug))
-        } else {
-          crumb.path = ""
-        }
+        crumb.path = ""
      }

      return crumb
    })

-    if (baseMeta && options.showCurrentPage) {
-      crumbs.push({
-        displayName: baseMeta.currentView.replaceAll("-", " "),
-        path: "",
-      })
-    }
-
    if (!options.showCurrentPage) {
      crumbs.pop()
    }
--- a/quartz/components/index.ts
+++ b/quartz/components/index.ts
@@ -1,8 +1,6 @@
 import Content from "./pages/Content"
 import TagContent from "./pages/TagContent"
 import FolderContent from "./pages/FolderContent"
-import BaseContent from "./pages/BaseContent"
-import BaseViewSelector from "./BaseViewSelector"
 import NotFound from "./pages/404"
 import ArticleTitle from "./ArticleTitle"
 import Darkmode from "./Darkmode"
@@ -31,8 +29,6 @@ export {
  Content,
  TagContent,
  FolderContent,
-  BaseContent,
-  BaseViewSelector,
  Darkmode,
  ReaderMode,
  Head,
--- a/quartz/components/pages/BaseContent.tsx
+++ b/quartz/components/pages/BaseContent.tsx
@@ -1,20 +0,0 @@
-import { QuartzComponent, QuartzComponentConstructor, QuartzComponentProps } from "../types"
-import style from "../styles/basePage.scss"
-import { htmlToJsx } from "../../util/jsx"
-
-export default (() => {
-  const BaseContent: QuartzComponent = (props: QuartzComponentProps) => {
-    const { fileData, tree } = props
-
-    return (
-      <div class="popover-hint">
-        <article class={["base-content", ...(fileData.frontmatter?.cssclasses ?? [])].join(" ")}>
-          {htmlToJsx(fileData.filePath!, fileData.basesRenderedTree ?? tree)}
-        </article>
-      </div>
-    )
-  }
-
-  BaseContent.css = style
-  return BaseContent
-}) satisfies QuartzComponentConstructor
--- a/quartz/components/scripts/base-view-selector.inline.ts
+++ b/quartz/components/scripts/base-view-selector.inline.ts
@@ -1,144 +0,0 @@
-let documentClickHandler: ((e: MouseEvent) => void) | null = null
-
-function setupBaseViewSelector() {
-  const selectors = document.querySelectorAll("[data-base-view-selector]")
-
-  if (selectors.length === 0) return
-
-  if (!documentClickHandler) {
-    documentClickHandler = (e: MouseEvent) => {
-      document.querySelectorAll("[data-base-view-selector]").forEach((selector) => {
-        if (selector.contains(e.target as Node)) return
-        const trigger = selector.querySelector(".text-icon-button") as HTMLElement | null
-        if (trigger?.getAttribute("aria-expanded") === "true") {
-          selector.dispatchEvent(new CustomEvent("close-dropdown"))
-        }
-      })
-    }
-    document.addEventListener("click", documentClickHandler)
-    window.addCleanup(() => {
-      if (documentClickHandler) {
-        document.removeEventListener("click", documentClickHandler)
-        documentClickHandler = null
-      }
-    })
-  }
-
-  selectors.forEach((selector) => {
-    if (selector.hasAttribute("data-initialized")) return
-    selector.setAttribute("data-initialized", "true")
-
-    const triggerEl = selector.querySelector(".text-icon-button") as HTMLElement | null
-    const searchInputEl = selector.querySelector("[data-search-input]") as HTMLInputElement | null
-    const clearButtonEl = selector.querySelector("[data-clear-search]") as HTMLElement | null
-    const viewListEl = selector.querySelector("[data-view-list]") as HTMLElement | null
-
-    if (!triggerEl || !searchInputEl || !clearButtonEl || !viewListEl) return
-
-    const trigger = triggerEl
-    const searchInput = searchInputEl
-    const clearButton = clearButtonEl
-    const viewList = viewListEl
-
-    function toggleDropdown() {
-      if (trigger.getAttribute("aria-expanded") === "true") {
-        closeDropdown()
-        return
-      }
-      openDropdown()
-    }
-
-    function openDropdown() {
-      trigger.setAttribute("aria-expanded", "true")
-      trigger.classList.add("has-active-menu")
-      setTimeout(() => searchInput.focus(), 10)
-    }
-
-    function closeDropdown() {
-      trigger.setAttribute("aria-expanded", "false")
-      trigger.classList.remove("has-active-menu")
-      searchInput.value = ""
-      clearButton.hidden = true
-      filterViews("")
-    }
-
-    function filterViews(query: string) {
-      const items = viewList.querySelectorAll<HTMLElement>(".bases-toolbar-menu-item")
-      const lowerQuery = query.toLowerCase()
-
-      items.forEach((item) => {
-        const viewName = (item.getAttribute("data-view-name") || "").toLowerCase()
-        const viewType = (item.getAttribute("data-view-type") || "").toLowerCase()
-        const matches = viewName.includes(lowerQuery) || viewType.includes(lowerQuery)
-        item.style.display = matches ? "" : "none"
-      })
-    }
-
-    function handleSearchInput() {
-      const query = searchInput.value
-      filterViews(query)
-      clearButton.hidden = query.length === 0
-    }
-
-    function clearSearch() {
-      searchInput.value = ""
-      clearButton.hidden = true
-      filterViews("")
-      searchInput.focus()
-    }
-
-    const handleTriggerClick = (e: MouseEvent) => {
-      e.stopPropagation()
-      toggleDropdown()
-    }
-
-    const handleTriggerKeydown = (e: KeyboardEvent) => {
-      if (e.key === "Enter" || e.key === " ") {
-        e.preventDefault()
-        toggleDropdown()
-      }
-    }
-
-    const handleSearchKeydown = (e: KeyboardEvent) => {
-      if (e.key === "Escape") {
-        if (searchInput.value) {
-          clearSearch()
-        } else {
-          closeDropdown()
-        }
-      }
-    }
-
-    const handleClearClick = (e: MouseEvent) => {
-      e.stopPropagation()
-      clearSearch()
-    }
-
-    trigger.addEventListener("click", handleTriggerClick)
-    trigger.addEventListener("keydown", handleTriggerKeydown)
-    searchInput.addEventListener("input", handleSearchInput)
-    searchInput.addEventListener("keydown", handleSearchKeydown)
-    clearButton.addEventListener("click", handleClearClick)
-
-    const viewLinks = viewList.querySelectorAll(".bases-toolbar-menu-item")
-    viewLinks.forEach((link) => {
-      link.addEventListener("click", closeDropdown)
-      window.addCleanup(() => link.removeEventListener("click", closeDropdown))
-    })
-
-    selector.addEventListener("close-dropdown", closeDropdown)
-
-    window.addCleanup(() => {
-      trigger.removeEventListener("click", handleTriggerClick)
-      trigger.removeEventListener("keydown", handleTriggerKeydown)
-      searchInput.removeEventListener("input", handleSearchInput)
-      searchInput.removeEventListener("keydown", handleSearchKeydown)
-      clearButton.removeEventListener("click", handleClearClick)
-      selector.removeEventListener("close-dropdown", closeDropdown)
-      selector.removeAttribute("data-initialized")
-      closeDropdown()
-    })
-  })
-}
-
-document.addEventListener("nav", setupBaseViewSelector)
--- a/quartz/components/styles/basePage.scss
+++ b/quartz/components/styles/basePage.scss
@@ -1,299 +0,0 @@
-.base-content {
-  width: 100%;
-}
-
-.base-view {
-  width: 100%;
-  overflow-x: auto;
-}
-
-.base-table {
-  width: 100%;
-  border-collapse: collapse;
-  font-size: 0.875rem;
-
-  th,
-  td {
-    padding: 0.5rem 0.75rem;
-    text-align: left;
-    border-bottom: 1px solid var(--lightgray);
-  }
-
-  th {
-    font-weight: 600;
-    color: var(--darkgray);
-    background: var(--light);
-    position: sticky;
-    top: 0;
-  }
-
-  tbody tr:hover {
-    background: var(--light);
-  }
-
-  a.internal {
-    color: var(--secondary);
-    text-decoration: none;
-
-    &:hover {
-      text-decoration: underline;
-    }
-  }
-}
-
-.base-group-header td {
-  font-weight: 600;
-  background: var(--light);
-  color: var(--dark);
-  padding-top: 1rem;
-}
-
-.base-summary-row {
-  background: var(--light);
-  font-weight: 500;
-
-  .base-summary-cell {
-    border-top: 2px solid var(--lightgray);
-    color: var(--darkgray);
-  }
-}
-
-.base-checkbox {
-  pointer-events: none;
-  width: 1rem;
-  height: 1rem;
-  accent-color: var(--secondary);
-}
-
-.base-list {
-  list-style: none;
-  padding: 0;
-  margin: 0;
-
-  li {
-    padding: 0.375rem 0;
-    border-bottom: 1px solid var(--lightgray);
-
-    &:last-child {
-      border-bottom: none;
-    }
-  }
-
-  a.internal {
-    color: var(--secondary);
-    text-decoration: none;
-
-    &:hover {
-      text-decoration: underline;
-    }
-  }
-}
-
-.base-list-container {
-  display: flex;
-  flex-direction: column;
-  gap: 1.5rem;
-}
-
-.base-list-group {
-  .base-list-group-header {
-    font-size: 1rem;
-    font-weight: 600;
-    margin-bottom: 0.5rem;
-    color: var(--dark);
-  }
-}
-
-.base-list-nested {
-  list-style: none;
-  padding-left: 1rem;
-  margin-top: 0.25rem;
-  font-size: 0.8125rem;
-  color: var(--darkgray);
-}
-
-.base-list-meta-label {
-  font-weight: 500;
-}
-
-.base-card-grid {
-  --base-card-min: 200px;
-  --base-card-aspect: 1.4;
-
-  display: grid;
-  grid-template-columns: repeat(auto-fill, minmax(var(--base-card-min), 1fr));
-  gap: 1rem;
-}
-
-.base-card-container {
-  display: flex;
-  flex-direction: column;
-  gap: 1.5rem;
-}
-
-.base-card-group {
-  .base-card-group-header {
-    font-size: 1rem;
-    font-weight: 600;
-    margin-bottom: 0.75rem;
-    color: var(--dark);
-  }
-}
-
-.base-card {
-  display: flex;
-  flex-direction: column;
-  border: 1px solid var(--lightgray);
-  border-radius: 8px;
-  overflow: hidden;
-  background: var(--light);
-  transition: box-shadow 0.15s ease;
-
-  &:hover {
-    box-shadow: 0 4px 12px rgba(0, 0, 0, 0.1);
-  }
-}
-
-.base-card-image-link {
-  display: block;
-  aspect-ratio: var(--base-card-aspect);
-  background-position: center;
-  background-repeat: no-repeat;
-}
-
-.base-card-content {
-  padding: 0.75rem;
-  display: flex;
-  flex-direction: column;
-  gap: 0.5rem;
-}
-
-.base-card-title-link {
-  text-decoration: none;
-  color: inherit;
-
-  &:hover .base-card-title {
-    color: var(--secondary);
-  }
-}
-
-.base-card-title {
-  font-size: 0.9375rem;
-  font-weight: 600;
-  margin: 0;
-  line-height: 1.3;
-  transition: color 0.15s ease;
-}
-
-.base-card-meta {
-  display: flex;
-  flex-direction: column;
-  gap: 0.25rem;
-  font-size: 0.8125rem;
-  color: var(--darkgray);
-}
-
-.base-card-meta-item {
-  display: flex;
-  gap: 0.25rem;
-}
-
-.base-card-meta-label {
-  font-weight: 500;
-
-  &::after {
-    content: ":";
-  }
-}
-
-.base-calendar-container {
-  display: flex;
-  flex-direction: column;
-  gap: 1.5rem;
-}
-
-.base-calendar-group {
-  .base-calendar-group-header {
-    font-size: 0.9375rem;
-    font-weight: 600;
-    margin-bottom: 0.5rem;
-    color: var(--dark);
-    font-variant-numeric: tabular-nums;
-  }
-}
-
-.base-map {
-  width: 100%;
-  min-height: 400px;
-  background: var(--light);
-  border: 1px solid var(--lightgray);
-  border-radius: 8px;
-  display: flex;
-  align-items: center;
-  justify-content: center;
-  color: var(--darkgray);
-
-  &::before {
-    content: "Map view requires client-side JavaScript";
-    font-size: 0.875rem;
-  }
-}
-
-.base-diagnostics {
-  background: #fff3cd;
-  border: 1px solid #ffc107;
-  border-radius: 4px;
-  padding: 1rem;
-  margin-bottom: 1rem;
-  font-size: 0.875rem;
-}
-
-.base-diagnostics-title {
-  font-weight: 600;
-  margin-bottom: 0.5rem;
-  color: #856404;
-}
-
-.base-diagnostics-meta {
-  display: flex;
-  gap: 0.5rem;
-  margin-bottom: 0.75rem;
-  color: #856404;
-}
-
-.base-diagnostics-page {
-  font-family: var(--codeFont);
-}
-
-.base-diagnostics-list {
-  list-style: none;
-  padding: 0;
-  margin: 0;
-  display: flex;
-  flex-direction: column;
-  gap: 0.5rem;
-}
-
-.base-diagnostics-item {
-  background: white;
-  padding: 0.5rem;
-  border-radius: 4px;
-}
-
-.base-diagnostics-label {
-  font-weight: 500;
-  color: #856404;
-}
-
-.base-diagnostics-message {
-  color: #664d03;
-  margin: 0.25rem 0;
-}
-
-.base-diagnostics-source {
-  display: block;
-  font-size: 0.8125rem;
-  color: #6c757d;
-  white-space: pre-wrap;
-  word-break: break-all;
-}
--- a/quartz/components/styles/baseViewSelector.scss
+++ b/quartz/components/styles/baseViewSelector.scss
@@ -1,275 +0,0 @@
-@use "../../styles/variables.scss" as *;
-
-.bases-toolbar {
-  position: relative;
-  display: inline-block;
-  margin: 1rem 0;
-  font-family: var(--bodyFont);
-
-  .bases-toolbar-item {
-    display: inline-block;
-    position: relative;
-
-    &.bases-toolbar-views-menu {
-      .text-icon-button {
-        display: flex;
-        align-items: center;
-        gap: 0.375rem;
-        padding: 0.375rem 0.75rem;
-        background: var(--light);
-        border: 1px solid var(--lightgray);
-        border-radius: 6px;
-        color: var(--darkgray);
-        font-size: 0.875rem;
-        font-weight: 500;
-        cursor: pointer;
-        transition: all 0.15s ease;
-        user-select: none;
-
-        &:hover {
-          background: var(--highlight);
-          border-color: var(--gray);
-        }
-
-        &.has-active-menu {
-          border-color: var(--secondary);
-          background: var(--highlight);
-        }
-
-        .text-button-icon {
-          display: flex;
-          align-items: center;
-          justify-content: center;
-          width: 16px;
-          height: 16px;
-          color: var(--gray);
-          flex-shrink: 0;
-
-          svg {
-            width: 16px;
-            height: 16px;
-          }
-
-          &.mod-aux {
-            opacity: 0.7;
-          }
-        }
-
-        .text-button-label {
-          font-size: 0.875rem;
-          color: var(--dark);
-          font-weight: 500;
-        }
-      }
-    }
-  }
-
-  .menu-scroll {
-    position: absolute;
-    top: calc(100% + 0.5rem);
-    left: 0;
-    z-index: 100;
-    max-height: 400px;
-    background: var(--light);
-    border: 1px solid var(--lightgray);
-    border-radius: 8px;
-    box-shadow:
-      0 4px 6px -1px rgb(0 0 0 / 0.1),
-      0 2px 4px -2px rgb(0 0 0 / 0.1);
-    overflow: hidden;
-    min-width: 280px;
-    display: none;
-  }
-
-  &:has(.text-icon-button.has-active-menu) .menu-scroll {
-    display: block;
-  }
-
-  .bases-toolbar-menu-container {
-    display: flex;
-    flex-direction: column;
-    max-height: 400px;
-
-    .search-input-container {
-      position: relative;
-      padding: 0.5rem;
-      border-bottom: 1px solid var(--lightgray);
-
-      input[type="search"] {
-        width: 100%;
-        padding: 0.375rem 0.75rem;
-        padding-right: 2rem;
-        background: var(--light);
-        border: 1px solid var(--secondary);
-        border-radius: 6px;
-        font-size: 0.875rem;
-        color: var(--dark);
-        outline: none;
-        transition: box-shadow 0.15s ease;
-        font-family: var(--bodyFont);
-
-        &::placeholder {
-          color: var(--gray);
-          opacity: 0.7;
-        }
-
-        &:focus {
-          box-shadow: 0 0 0 2px var(--highlight);
-        }
-
-        &::-webkit-search-cancel-button {
-          display: none;
-        }
-      }
-
-      .search-input-clear-button {
-        position: absolute;
-        right: 1rem;
-        top: 50%;
-        transform: translateY(-50%);
-        display: flex;
-        align-items: center;
-        justify-content: center;
-        width: 20px;
-        height: 20px;
-        cursor: pointer;
-        opacity: 0.5;
-        transition: opacity 0.15s ease;
-        color: var(--gray);
-
-        &:hover {
-          opacity: 1;
-        }
-
-        &[hidden] {
-          display: none;
-        }
-
-        svg {
-          width: 14px;
-          height: 14px;
-        }
-      }
-    }
-
-    .bases-toolbar-items {
-      overflow-y: auto;
-      max-height: 340px;
-
-      &::-webkit-scrollbar {
-        width: 8px;
-      }
-
-      &::-webkit-scrollbar-track {
-        background: transparent;
-      }
-
-      &::-webkit-scrollbar-thumb {
-        background: var(--lightgray);
-        border-radius: 4px;
-
-        &:hover {
-          background: var(--gray);
-        }
-      }
-
-      .suggestion-group {
-        &[data-group="views"] {
-          padding: 0.25rem 0;
-          text-transform: lowercase;
-        }
-      }
-
-      .suggestion-item {
-        display: block;
-        text-decoration: none;
-        color: inherit;
-        cursor: pointer;
-
-        &.bases-toolbar-menu-item {
-          display: flex;
-          align-items: center;
-          justify-content: space-between;
-          padding: 0.5rem 0.75rem;
-          margin: 0 0.25rem;
-          border-radius: 4px;
-          transition: background 0.15s ease;
-
-          &:hover {
-            background: var(--lightgray);
-          }
-
-          &.mod-active {
-            font-weight: $semiBoldWeight;
-          }
-
-          &.is-selected {
-            .bases-toolbar-menu-item-info {
-              .bases-toolbar-menu-item-name {
-                font-weight: 600;
-                color: var(--secondary);
-              }
-            }
-          }
-
-          .bases-toolbar-menu-item-info {
-            display: flex;
-            align-items: center;
-            gap: 0.5rem;
-            flex: 1;
-
-            .bases-toolbar-menu-item-info-icon {
-              display: flex;
-              align-items: center;
-              justify-content: center;
-              width: 16px;
-              height: 16px;
-              color: var(--gray);
-              flex-shrink: 0;
-
-              svg {
-                width: 16px;
-                height: 16px;
-              }
-            }
-
-            .bases-toolbar-menu-item-name {
-              font-size: 0.875rem;
-              color: var(--dark);
-            }
-          }
-
-          .clickable-icon.bases-toolbar-menu-item-icon {
-            display: flex;
-            align-items: center;
-            justify-content: center;
-            width: 16px;
-            height: 16px;
-            opacity: 0;
-            transition: opacity 0.15s ease;
-            color: var(--gray);
-            flex-shrink: 0;
-
-            svg {
-              width: 16px;
-              height: 16px;
-            }
-          }
-
-          &:hover .clickable-icon.bases-toolbar-menu-item-icon {
-            opacity: 0.5;
-          }
-        }
-      }
-    }
-  }
-}
-
-@media all and ($mobile) {
-  .bases-toolbar {
-    .menu-scroll {
-      min-width: 240px;
-      left: auto;
-    }
-  }
-}
--- a/quartz/plugins/emitters/assets.ts
+++ b/quartz/plugins/emitters/assets.ts
@@ -7,12 +7,8 @@ import { Argv } from "../../util/ctx"
 import { QuartzConfig } from "../../cfg"

 const filesToCopy = async (argv: Argv, cfg: QuartzConfig) => {
-  // glob all non MD/base files in content folder and copy it over
-  return await glob("**", argv.directory, [
-    "**/*.md",
-    "**/*.base",
-    ...cfg.configuration.ignorePatterns,
-  ])
+  // glob all non MD files in content folder and copy it over
+  return await glob("**", argv.directory, ["**/*.md", ...cfg.configuration.ignorePatterns])
 }

 const copyFile = async (argv: Argv, fp: FilePath) => {
@@ -41,7 +37,7 @@ export const Assets: QuartzEmitterPlugin = () => {
    async *partialEmit(ctx, _content, _resources, changeEvents) {
      for (const changeEvent of changeEvents) {
        const ext = path.extname(changeEvent.path)
-        if (ext === ".md" || ext === ".base") continue
+        if (ext === ".md") continue

        if (changeEvent.type === "add" || changeEvent.type === "change") {
          yield copyFile(ctx.argv, changeEvent.path)
--- a/quartz/plugins/emitters/basePage.tsx
+++ b/quartz/plugins/emitters/basePage.tsx
@@ -1,184 +0,0 @@
-import { QuartzEmitterPlugin } from "../types"
-import { QuartzComponentProps } from "../../components/types"
-import HeaderConstructor from "../../components/Header"
-import BodyConstructor from "../../components/Body"
-import { pageResources, renderPage } from "../../components/renderPage"
-import { ProcessedContent, QuartzPluginData } from "../vfile"
-import { FullPageLayout } from "../../cfg"
-import { pathToRoot } from "../../util/path"
-import { defaultListPageLayout, sharedPageComponents } from "../../../quartz.layout"
-import { BaseContent, BaseViewSelector } from "../../components"
-import { write } from "./helpers"
-import { BuildCtx } from "../../util/ctx"
-import { StaticResources } from "../../util/resources"
-import {
-  renderBaseViewsForFile,
-  RenderedBaseView,
-  BaseViewMeta,
-  BaseMetadata,
-} from "../../util/base/render"
-import { BaseFile } from "../../util/base/types"
-
-interface BasePageOptions extends FullPageLayout {}
-
-function isBaseFile(data: QuartzPluginData): boolean {
-  return Boolean(data.basesConfig && (data.basesConfig as BaseFile).views?.length > 0)
-}
-
-function getBaseFiles(content: ProcessedContent[]): ProcessedContent[] {
-  return content.filter(([_, file]) => isBaseFile(file.data))
-}
-
-async function processBasePage(
-  ctx: BuildCtx,
-  baseFileData: QuartzPluginData,
-  renderedView: RenderedBaseView,
-  allViews: BaseViewMeta[],
-  allFiles: QuartzPluginData[],
-  opts: FullPageLayout,
-  resources: StaticResources,
-) {
-  const slug = renderedView.slug
-  const cfg = ctx.cfg.configuration
-  const externalResources = pageResources(pathToRoot(slug), resources)
-
-  const viewFileData: QuartzPluginData = {
-    ...baseFileData,
-    slug,
-    frontmatter: {
-      ...baseFileData.frontmatter,
-      title: renderedView.view.name,
-    },
-    basesRenderedTree: renderedView.tree,
-    basesAllViews: allViews,
-    basesCurrentView: renderedView.view.name,
-    basesMetadata: {
-      baseSlug: baseFileData.slug!,
-      currentView: renderedView.view.name,
-      allViews,
-    },
-  }
-
-  const componentData: QuartzComponentProps = {
-    ctx,
-    fileData: viewFileData,
-    externalResources,
-    cfg,
-    children: [],
-    tree: renderedView.tree,
-    allFiles,
-  }
-
-  const content = renderPage(cfg, slug, componentData, opts, externalResources)
-  return write({
-    ctx,
-    content,
-    slug,
-    ext: ".html",
-  })
-}
-
-export const BasePage: QuartzEmitterPlugin<Partial<BasePageOptions>> = (userOpts) => {
-  const baseOpts: FullPageLayout = {
-    ...sharedPageComponents,
-    ...defaultListPageLayout,
-    pageBody: BaseContent(),
-    ...userOpts,
-  }
-
-  const opts: FullPageLayout = {
-    ...baseOpts,
-    beforeBody: [
-      ...baseOpts.beforeBody.filter((component) => component.name !== "ArticleTitle"),
-      BaseViewSelector(),
-    ],
-  }
-
-  const { head: Head, header, beforeBody, pageBody, afterBody, left, right, footer: Footer } = opts
-  const Header = HeaderConstructor()
-  const Body = BodyConstructor()
-
-  return {
-    name: "BasePage",
-    getQuartzComponents() {
-      return [
-        Head,
-        Header,
-        Body,
-        ...header,
-        ...beforeBody,
-        pageBody,
-        ...afterBody,
-        ...left,
-        ...right,
-        Footer,
-      ]
-    },
-    async *emit(ctx, content, resources) {
-      const allFiles = content.map((c) => c[1].data)
-      const baseFiles = getBaseFiles(content)
-
-      for (const [_, file] of baseFiles) {
-        const baseFileData = file.data
-        const { views, allViews } = renderBaseViewsForFile(baseFileData, allFiles)
-
-        for (const renderedView of views) {
-          yield processBasePage(
-            ctx,
-            baseFileData,
-            renderedView,
-            allViews,
-            allFiles,
-            opts,
-            resources,
-          )
-        }
-      }
-    },
-    async *partialEmit(ctx, content, resources, changeEvents) {
-      const allFiles = content.map((c) => c[1].data)
-      const baseFiles = getBaseFiles(content)
-
-      const affectedBaseSlugs = new Set<string>()
-
-      for (const event of changeEvents) {
-        if (!event.file) continue
-        const slug = event.file.data.slug
-
-        if (slug && isBaseFile(event.file.data)) {
-          affectedBaseSlugs.add(slug)
-        }
-      }
-
-      for (const [_, file] of baseFiles) {
-        const baseFileData = file.data
-        const baseSlug = baseFileData.slug
-
-        if (!baseSlug || !affectedBaseSlugs.has(baseSlug)) continue
-
-        const { views, allViews } = renderBaseViewsForFile(baseFileData, allFiles)
-
-        for (const renderedView of views) {
-          yield processBasePage(
-            ctx,
-            baseFileData,
-            renderedView,
-            allViews,
-            allFiles,
-            opts,
-            resources,
-          )
-        }
-      }
-    },
-  }
-}
-
-declare module "vfile" {
-  interface DataMap {
-    basesRenderedTree?: import("hast").Root
-    basesAllViews?: BaseViewMeta[]
-    basesCurrentView?: string
-    basesMetadata?: BaseMetadata
-  }
-}
--- a/quartz/plugins/emitters/contentPage.tsx
+++ b/quartz/plugins/emitters/contentPage.tsx
@@ -83,8 +83,6 @@ export const ContentPage: QuartzEmitterPlugin<Partial<FullPageLayout>> = (userOp
          containsIndex = true
        }

-        if (file.data.filePath!.endsWith(".base")) continue
-
        // only process home page, non-tag pages, and non-index pages
        if (slug.endsWith("/index") || slug.startsWith("tags/")) continue
        yield processContent(ctx, tree, file.data, allFiles, opts, resources)
@@ -114,7 +112,6 @@ export const ContentPage: QuartzEmitterPlugin<Partial<FullPageLayout>> = (userOp
      for (const [tree, file] of content) {
        const slug = file.data.slug!
        if (!changedSlugs.has(slug)) continue
-        if (file.data.filePath!.endsWith(".base")) continue
        if (slug.endsWith("/index") || slug.startsWith("tags/")) continue

        yield processContent(ctx, tree, file.data, allFiles, opts, resources)
--- a/quartz/plugins/emitters/index.ts
+++ b/quartz/plugins/emitters/index.ts
@@ -10,4 +10,3 @@ export { ComponentResources } from "./componentResources"
 export { NotFoundPage } from "./404"
 export { CNAME } from "./cname"
 export { CustomOgImages } from "./ogImage"
-export { BasePage } from "./basePage"
--- a/quartz/plugins/emitters/static.ts
+++ b/quartz/plugins/emitters/static.ts
@@ -7,6 +7,7 @@ import { dirname } from "path"
 export const Static: QuartzEmitterPlugin = () => ({
  name: "Static",
  async *emit({ argv, cfg }) {
+    // Copy Quartz's own internal static assets (quartz/static/) → output/static/
    const staticPath = joinSegments(QUARTZ, "static")
    const fps = await glob("**", staticPath, cfg.configuration.ignorePatterns)
    const outputStaticPath = joinSegments(argv.output, "static")
@@ -18,6 +19,21 @@ export const Static: QuartzEmitterPlugin = () => ({
      await fs.promises.copyFile(src, dest)
      yield dest
    }
+
+    // Copy user-facing static assets (static/) → output/ preserving paths.
+    // This mirrors Hugo's convention: static/ox-hugo/foo.png is served at /ox-hugo/foo.png,
+    // which matches the src="/ox-hugo/..." paths that ox-hugo writes into exported markdown.
+    const userStaticPath = "static"
+    if (fs.existsSync(userStaticPath)) {
+      const userFps = await glob("**", userStaticPath, cfg.configuration.ignorePatterns, false)
+      for (const fp of userFps) {
+        const src = joinSegments(userStaticPath, fp) as FilePath
+        const dest = joinSegments(argv.output, fp) as FilePath
+        await fs.promises.mkdir(dirname(dest), { recursive: true })
+        await fs.promises.copyFile(src, dest)
+        yield dest
+      }
+    }
  },
  async *partialEmit() {},
 })
--- a/quartz/plugins/transformers/bases.ts
+++ b/quartz/plugins/transformers/bases.ts
@@ -1,521 +0,0 @@
-import * as yaml from "js-yaml"
-import { QuartzTransformerPlugin } from "../types"
-import { FilePath, getFileExtension } from "../../util/path"
-import {
-  BaseFile,
-  BaseView,
-  BaseFileFilter,
-  parseViews,
-  parseViewSummaries,
-  BUILTIN_SUMMARY_TYPES,
-  BuiltinSummaryType,
-} from "../../util/base/types"
-import {
-  parseExpressionSource,
-  compileExpression,
-  buildPropertyExpressionSource,
-  ProgramIR,
-  BasesExpressions,
-  BaseExpressionDiagnostic,
-  Span,
-} from "../../util/base/compiler"
-
-export interface BasesOptions {
-  /** Whether to emit diagnostics as warnings during build */
-  emitWarnings: boolean
-}
-
-const defaultOptions: BasesOptions = {
-  emitWarnings: true,
-}
-
-type FilterStructure =
-  | string
-  | { and?: FilterStructure[]; or?: FilterStructure[]; not?: FilterStructure[] }
-
-function compileFilterStructure(
-  filter: FilterStructure | undefined,
-  file: string,
-  diagnostics: BaseExpressionDiagnostic[],
-  context: string,
-): ProgramIR | undefined {
-  if (!filter) return undefined
-
-  if (typeof filter === "string") {
-    const result = parseExpressionSource(filter, file)
-    if (result.diagnostics.length > 0) {
-      for (const diag of result.diagnostics) {
-        diagnostics.push({
-          kind: diag.kind as "lex" | "parse" | "runtime",
-          message: diag.message,
-          span: diag.span,
-          context,
-          source: filter,
-        })
-      }
-    }
-    if (!result.program.body) return undefined
-    return compileExpression(result.program.body)
-  }
-
-  const compileParts = (
-    parts: FilterStructure[],
-    combiner: "&&" | "||",
-    negate: boolean,
-  ): ProgramIR | undefined => {
-    const compiled: ProgramIR[] = []
-    for (const part of parts) {
-      const partIR = compileFilterStructure(part, file, diagnostics, context)
-      if (partIR) compiled.push(partIR)
-    }
-    if (compiled.length === 0) return undefined
-    if (compiled.length === 1) {
-      if (negate) {
-        return wrapWithNot(compiled[0])
-      }
-      return compiled[0]
-    }
-
-    let result = compiled[0]
-    for (let i = 1; i < compiled.length; i++) {
-      result = combineWithLogical(result, compiled[i], combiner, negate)
-    }
-    return result
-  }
-
-  if (filter.and && filter.and.length > 0) {
-    return compileParts(filter.and, "&&", false)
-  }
-  if (filter.or && filter.or.length > 0) {
-    return compileParts(filter.or, "||", false)
-  }
-  if (filter.not && filter.not.length > 0) {
-    return compileParts(filter.not, "&&", true)
-  }
-
-  return undefined
-}
-
-function wrapWithNot(ir: ProgramIR): ProgramIR {
-  const span = ir.span
-  return {
-    instructions: [
-      ...ir.instructions,
-      { op: "to_bool" as const, span },
-      { op: "unary" as const, operator: "!" as const, span },
-    ],
-    span,
-  }
-}
-
-function combineWithLogical(
-  left: ProgramIR,
-  right: ProgramIR,
-  operator: "&&" | "||",
-  negateRight: boolean,
-): ProgramIR {
-  const span: Span = {
-    start: left.span.start,
-    end: right.span.end,
-    file: left.span.file,
-  }
-
-  const rightIR = negateRight ? wrapWithNot(right) : right
-
-  if (operator === "&&") {
-    const jumpIfFalseIndex = left.instructions.length + 1
-    const jumpIndex = jumpIfFalseIndex + rightIR.instructions.length + 2
-    return {
-      instructions: [
-        ...left.instructions,
-        { op: "jump_if_false" as const, target: jumpIndex, span },
-        ...rightIR.instructions,
-        { op: "to_bool" as const, span },
-        { op: "jump" as const, target: jumpIndex + 1, span },
-        {
-          op: "const" as const,
-          literal: { type: "Literal" as const, kind: "boolean" as const, value: false, span },
-          span,
-        },
-      ],
-      span,
-    }
-  } else {
-    const jumpIfTrueIndex = left.instructions.length + 1
-    const jumpIndex = jumpIfTrueIndex + rightIR.instructions.length + 2
-    return {
-      instructions: [
-        ...left.instructions,
-        { op: "jump_if_true" as const, target: jumpIndex, span },
-        ...rightIR.instructions,
-        { op: "to_bool" as const, span },
-        { op: "jump" as const, target: jumpIndex + 1, span },
-        {
-          op: "const" as const,
-          literal: { type: "Literal" as const, kind: "boolean" as const, value: true, span },
-          span,
-        },
-      ],
-      span,
-    }
-  }
-}
-
-function collectPropertiesFromViews(views: BaseView[]): Set<string> {
-  const properties = new Set<string>()
-  for (const view of views) {
-    if (view.order) {
-      for (const prop of view.order) {
-        properties.add(prop)
-      }
-    }
-    if (view.groupBy) {
-      const groupProp = typeof view.groupBy === "string" ? view.groupBy : view.groupBy.property
-      properties.add(groupProp)
-    }
-    if (view.sort) {
-      for (const sortConfig of view.sort) {
-        properties.add(sortConfig.property)
-      }
-    }
-    if (view.image) properties.add(view.image)
-    if (view.date) properties.add(view.date)
-    if (view.dateField) properties.add(view.dateField)
-    if (view.dateProperty) properties.add(view.dateProperty)
-    if (view.coordinates) properties.add(view.coordinates)
-    if (view.markerIcon) properties.add(view.markerIcon)
-    if (view.markerColor) properties.add(view.markerColor)
-  }
-  return properties
-}
-
-function compilePropertyExpressions(
-  properties: Set<string>,
-  file: string,
-  diagnostics: BaseExpressionDiagnostic[],
-): Record<string, ProgramIR> {
-  const expressions: Record<string, ProgramIR> = {}
-
-  for (const property of properties) {
-    const source = buildPropertyExpressionSource(property)
-    if (!source) continue
-
-    const result = parseExpressionSource(source, file)
-    if (result.diagnostics.length > 0) {
-      for (const diag of result.diagnostics) {
-        diagnostics.push({
-          kind: diag.kind as "lex" | "parse" | "runtime",
-          message: diag.message,
-          span: diag.span,
-          context: `property.${property}`,
-          source,
-        })
-      }
-    }
-    if (result.program.body) {
-      expressions[property] = compileExpression(result.program.body)
-    }
-  }
-
-  return expressions
-}
-
-function compileFormulas(
-  formulas: Record<string, string> | undefined,
-  file: string,
-  diagnostics: BaseExpressionDiagnostic[],
-): Record<string, ProgramIR> {
-  if (!formulas) return {}
-
-  const compiled: Record<string, ProgramIR> = {}
-  for (const [name, source] of Object.entries(formulas)) {
-    const trimmed = source.trim()
-    if (!trimmed) continue
-
-    const result = parseExpressionSource(trimmed, file)
-    if (result.diagnostics.length > 0) {
-      for (const diag of result.diagnostics) {
-        diagnostics.push({
-          kind: diag.kind as "lex" | "parse" | "runtime",
-          message: diag.message,
-          span: diag.span,
-          context: `formulas.${name}`,
-          source: trimmed,
-        })
-      }
-    }
-    if (result.program.body) {
-      compiled[name] = compileExpression(result.program.body)
-    }
-  }
-
-  return compiled
-}
-
-function compileSummaries(
-  summaries: Record<string, string> | undefined,
-  file: string,
-  diagnostics: BaseExpressionDiagnostic[],
-): Record<string, ProgramIR> {
-  if (!summaries) return {}
-
-  const compiled: Record<string, ProgramIR> = {}
-  for (const [name, source] of Object.entries(summaries)) {
-    const trimmed = source.trim()
-    if (!trimmed) continue
-
-    const normalized = trimmed.toLowerCase()
-    if (BUILTIN_SUMMARY_TYPES.includes(normalized as BuiltinSummaryType)) {
-      continue
-    }
-
-    const result = parseExpressionSource(trimmed, file)
-    if (result.diagnostics.length > 0) {
-      for (const diag of result.diagnostics) {
-        diagnostics.push({
-          kind: diag.kind as "lex" | "parse" | "runtime",
-          message: diag.message,
-          span: diag.span,
-          context: `summaries.${name}`,
-          source: trimmed,
-        })
-      }
-    }
-    if (result.program.body) {
-      compiled[name] = compileExpression(result.program.body)
-    }
-  }
-
-  return compiled
-}
-
-function compileViewSummaries(
-  views: BaseView[],
-  topLevelSummaries: Record<string, string> | undefined,
-  file: string,
-  diagnostics: BaseExpressionDiagnostic[],
-): Record<string, Record<string, ProgramIR>> {
-  const result: Record<string, Record<string, ProgramIR>> = {}
-
-  for (let i = 0; i < views.length; i++) {
-    const view = views[i]
-    if (!view.summaries) continue
-
-    const viewSummaryConfig = parseViewSummaries(
-      view.summaries as Record<string, string>,
-      topLevelSummaries,
-    )
-    if (!viewSummaryConfig?.columns) continue
-
-    const viewExpressions: Record<string, ProgramIR> = {}
-    for (const [column, def] of Object.entries(viewSummaryConfig.columns)) {
-      if (def.type !== "formula" || !def.expression) continue
-
-      const parseResult = parseExpressionSource(def.expression, file)
-      if (parseResult.diagnostics.length > 0) {
-        for (const diag of parseResult.diagnostics) {
-          diagnostics.push({
-            kind: diag.kind as "lex" | "parse" | "runtime",
-            message: diag.message,
-            span: diag.span,
-            context: `views[${i}].summaries.${column}`,
-            source: def.expression,
-          })
-        }
-      }
-      if (parseResult.program.body) {
-        viewExpressions[column] = compileExpression(parseResult.program.body)
-      }
-    }
-
-    if (Object.keys(viewExpressions).length > 0) {
-      result[String(i)] = viewExpressions
-    }
-  }
-
-  return result
-}
-
-export const ObsidianBases: QuartzTransformerPlugin<Partial<BasesOptions>> = (userOpts) => {
-  const opts = { ...defaultOptions, ...userOpts }
-
-  return {
-    name: "ObsidianBases",
-    textTransform(_ctx, src) {
-      return src
-    },
-    markdownPlugins(_ctx) {
-      return [
-        () => {
-          return (_tree, file) => {
-            const filePath = file.data.filePath as FilePath | undefined
-            if (!filePath) return
-
-            const ext = getFileExtension(filePath)
-            if (ext !== ".base") return
-
-            const content = file.value.toString()
-            if (!content.trim()) return
-
-            const diagnostics: BaseExpressionDiagnostic[] = []
-            const filePathStr = filePath
-
-            try {
-              const parsed = yaml.load(content, { schema: yaml.JSON_SCHEMA }) as Record<
-                string,
-                unknown
-              >
-              if (!parsed || typeof parsed !== "object") {
-                diagnostics.push({
-                  kind: "parse",
-                  message: "Base file must contain a valid YAML object",
-                  span: {
-                    start: { offset: 0, line: 1, column: 1 },
-                    end: { offset: 0, line: 1, column: 1 },
-                    file: filePathStr,
-                  },
-                  context: "root",
-                  source: content.slice(0, 100),
-                })
-                file.data.basesDiagnostics = diagnostics
-                return
-              }
-
-              const rawViews = parsed.views
-              if (!Array.isArray(rawViews) || rawViews.length === 0) {
-                diagnostics.push({
-                  kind: "parse",
-                  message: "Base file must have at least one view defined",
-                  span: {
-                    start: { offset: 0, line: 1, column: 1 },
-                    end: { offset: 0, line: 1, column: 1 },
-                    file: filePathStr,
-                  },
-                  context: "views",
-                  source: "views: []",
-                })
-                file.data.basesDiagnostics = diagnostics
-                return
-              }
-
-              const views = parseViews(rawViews)
-              const filters = parsed.filters as BaseFileFilter | undefined
-              const properties = parsed.properties as
-                | Record<string, { displayName?: string }>
-                | undefined
-              const summaries = parsed.summaries as Record<string, string> | undefined
-              const formulas = parsed.formulas as Record<string, string> | undefined
-
-              const baseConfig: BaseFile = {
-                filters,
-                views,
-                properties,
-                summaries,
-                formulas,
-              }
-
-              const compiledFilters = compileFilterStructure(
-                filters as FilterStructure | undefined,
-                filePathStr,
-                diagnostics,
-                "filters",
-              )
-
-              const viewFilters: Record<string, ProgramIR> = {}
-              for (let i = 0; i < views.length; i++) {
-                const view = views[i]
-                if (view.filters) {
-                  const compiled = compileFilterStructure(
-                    view.filters as FilterStructure,
-                    filePathStr,
-                    diagnostics,
-                    `views[${i}].filters`,
-                  )
-                  if (compiled) {
-                    viewFilters[String(i)] = compiled
-                  }
-                }
-              }
-
-              const compiledFormulas = compileFormulas(formulas, filePathStr, diagnostics)
-
-              const compiledSummaries = compileSummaries(summaries, filePathStr, diagnostics)
-              const compiledViewSummaries = compileViewSummaries(
-                views,
-                summaries,
-                filePathStr,
-                diagnostics,
-              )
-
-              const viewProperties = collectPropertiesFromViews(views)
-
-              for (const name of Object.keys(compiledFormulas)) {
-                viewProperties.add(`formula.${name}`)
-              }
-
-              const propertyExpressions = compilePropertyExpressions(
-                viewProperties,
-                filePathStr,
-                diagnostics,
-              )
-
-              const expressions: BasesExpressions = {
-                filters: compiledFilters,
-                viewFilters,
-                formulas: compiledFormulas,
-                summaries: compiledSummaries,
-                viewSummaries: compiledViewSummaries,
-                propertyExpressions,
-              }
-
-              file.data.basesConfig = baseConfig
-              file.data.basesExpressions = expressions
-              file.data.basesDiagnostics = diagnostics
-
-              const existingFrontmatter = (file.data.frontmatter ?? {}) as Record<string, unknown>
-              file.data.frontmatter = {
-                title: views[0]?.name ?? file.stem ?? "Base",
-                tags: ["base"],
-                ...existingFrontmatter,
-              }
-
-              if (opts.emitWarnings && diagnostics.length > 0) {
-                for (const diag of diagnostics) {
-                  console.warn(
-                    `[bases] ${filePathStr}:${diag.span.start.line}:${diag.span.start.column} - ${diag.message}`,
-                  )
-                }
-              }
-            } catch (err) {
-              const message = err instanceof Error ? err.message : String(err)
-              diagnostics.push({
-                kind: "parse",
-                message: `Failed to parse base file: ${message}`,
-                span: {
-                  start: { offset: 0, line: 1, column: 1 },
-                  end: { offset: 0, line: 1, column: 1 },
-                  file: filePathStr,
-                },
-                context: "root",
-                source: content.slice(0, 100),
-              })
-              file.data.basesDiagnostics = diagnostics
-
-              if (opts.emitWarnings) {
-                console.warn(`[bases] ${filePathStr}: ${message}`)
-              }
-            }
-          }
-        },
-      ]
-    },
-  }
-}
-
-declare module "vfile" {
-  interface DataMap {
-    basesConfig?: BaseFile
-    basesExpressions?: BasesExpressions
-    basesDiagnostics?: BaseExpressionDiagnostic[]
-  }
-}
--- a/quartz/plugins/transformers/index.ts
+++ b/quartz/plugins/transformers/index.ts
@@ -11,4 +11,3 @@ export { SyntaxHighlighting } from "./syntax"
 export { TableOfContents } from "./toc"
 export { HardLineBreaks } from "./linebreaks"
 export { RoamFlavoredMarkdown } from "./roam"
-export { ObsidianBases } from "./bases"
--- a/quartz/plugins/transformers/ofm.ts
+++ b/quartz/plugins/transformers/ofm.ts
@@ -289,11 +289,8 @@ export const ObsidianFlavoredMarkdown: QuartzTransformerPlugin<Partial<Options>>
                  }
                }

-                const isBaseFile = fp.endsWith(".base")
-                const basePath = isBaseFile ? fp.slice(0, -5) : fp
-                const url = isBaseFile
-                  ? basePath + (anchor ? `/${anchor.slice(1).replace(/\s+/g, "-")}` : "")
-                  : fp + anchor
+                // internal link
+                const url = fp + anchor

                return {
                  type: "link",
@@ -301,7 +298,7 @@ export const ObsidianFlavoredMarkdown: QuartzTransformerPlugin<Partial<Options>>
                  children: [
                    {
                      type: "text",
-                      value: alias ?? basePath,
+                      value: alias ?? fp,
                    },
                  ],
                }
--- a/quartz/plugins/transformers/oxhugofm.ts
+++ b/quartz/plugins/transformers/oxhugofm.ts
@@ -27,7 +27,10 @@ const defaultOptions: Options = {
 const relrefRegex = new RegExp(/\[([^\]]+)\]\(\{\{< relref "([^"]+)" >\}\}\)/, "g")
 const predefinedHeadingIdRegex = new RegExp(/(.*) {#(?:.*)}/, "g")
 const hugoShortcodeRegex = new RegExp(/{{(.*)}}/, "g")
-const figureTagRegex = new RegExp(/< ?figure src="(.*)" ?>/, "g")
+// Matches the full Hugo {{< figure src="..." ... >}} shortcode and captures src.
+// Must run before the generic shortcode stripper to avoid partial-match issues
+// with captions that contain HTML (e.g. <span class="figure-number">).
+const figureShortcodeRegex = new RegExp(/{{<\s*figure\b[^}]*\bsrc="([^"]*)"[^}]*>}}/, "g")
 // \\\\\( -> matches \\(
 // (.+?) -> Lazy match for capturing the equation
 // \\\\\) -> matches \\)
@@ -70,6 +73,14 @@ export const OxHugoFlavouredMarkdown: QuartzTransformerPlugin<Partial<Options>>
        })
      }

+      if (opts.replaceFigureWithMdImg) {
+        src = src.toString()
+        src = src.replaceAll(figureShortcodeRegex, (_value, ...capture) => {
+          const [imgSrc] = capture
+          return `![](${imgSrc})`
+        })
+      }
+
      if (opts.removeHugoShortcode) {
        src = src.toString()
        src = src.replaceAll(hugoShortcodeRegex, (_value, ...capture) => {
@@ -78,14 +89,6 @@ export const OxHugoFlavouredMarkdown: QuartzTransformerPlugin<Partial<Options>>
        })
      }

-      if (opts.replaceFigureWithMdImg) {
-        src = src.toString()
-        src = src.replaceAll(figureTagRegex, (_value, ...capture) => {
-          const [src] = capture
-          return `![](${src})`
-        })
-      }
-
      if (opts.replaceOrgLatex) {
        src = src.toString()
        src = src.replaceAll(inlineLatexRegex, (_value, ...capture) => {
--- a/quartz/processors/parse.ts
+++ b/quartz/processors/parse.ts
@@ -104,16 +104,12 @@ export function createFileParser(ctx: BuildCtx, fps: FilePath[]) {
        file.data.relativePath = path.posix.relative(argv.directory, file.path) as FilePath
        file.data.slug = slugifyFilePath(file.data.relativePath)

-        const isBaseFile = fp.endsWith(".base")
-        const ast: MDRoot = isBaseFile ? { type: "root", children: [] } : processor.parse(file)
-
+        const ast = processor.parse(file)
        const newAst = await processor.run(ast, file)
        res.push([newAst, file])

        if (argv.verbose) {
-          console.log(
-            `[${isBaseFile ? "base" : "markdown"}] ${fp} -> ${file.data.slug} (${perf.timeSince()})`,
-          )
+          console.log(`[markdown] ${fp} -> ${file.data.slug} (${perf.timeSince()})`)
        }
      } catch (err) {
        trace(`\nFailed to process markdown \`${fp}\``, err as Error)
--- a/quartz/util/base/README.md
+++ b/quartz/util/base/README.md
@@ -1,92 +0,0 @@
-# bases compiler + runtime (quartz implementation)
-
-status: active
-last updated: 2026-01-28
-
-this directory contains the obsidian bases compiler, interpreter, and runtime helpers used by quartz to render `.base` files. it is designed to match obsidian bases syntax and semantics with deterministic evaluation and consistent diagnostics.
-
-You can test it out with any of the base file in my vault here:
-
-```bash
-npx tsx quartz/util/base/inspect-base.ts docs/navigation.base > /tmp/ast-ir.json
-
-jq '.expressions[] | {context, kind, source, ast}' /tmp/ast-ir.json
-jq '.expressions[] | {context, kind, ir}' /tmp/ast-ir.json
-```
-
-## scope
-
- parse base expressions (filters, formulas, summaries, property expressions)
- compile expressions to bytecode ir
- interpret bytecode with a deterministic stack vm
- resolve file, note, formula, and property values
- render views (table, list, cards/gallery, board, calendar, map)
- surface parse and runtime diagnostics in base output
-
-## architecture (pipeline)
-
-1. parse `.base` yaml (plugin: `quartz/plugins/transformers/bases.ts`)
-2. parse expressions into ast (`compiler/parser.ts`)
-3. compile ast to ir (`compiler/ir.ts`)
-4. evaluate ir per row with caches (`compiler/interpreter.ts`)
-5. render views and diagnostics (`render.ts`)
-
-## modules
-
- `compiler/lexer.ts`: tokenizer with span tracking and regex support
- `compiler/parser.ts`: pratt parser for expression grammar and error recovery
- `compiler/ir.ts`: bytecode instruction set + compiler
- `compiler/interpreter.ts`: stack vm, value model, coercions, methods, functions
- `compiler/diagnostics.ts`: diagnostics types and helpers
- `compiler/schema.ts`: summary config schema and builtins
- `compiler/properties.ts`: property expression builder for columns and config keys
- `render.ts`: view rendering and diagnostics output
- `query.ts`: summaries and view summary helpers
- `types.ts`: base config types and yaml parsing helpers
-
-## value model (runtime)
-
-runtime values are tagged unions with explicit kinds:
-
- null, boolean, number, string
- date, duration
- list, object
- file, link
- regex, html, icon, image
-
-coercions are permissive to match obsidian behavior. comparisons prefer type-aware equality (links resolve to files when possible, dates compare by time, etc), with fallbacks when resolution fails.
-
-## expression features (spec parity)
-
- operators: `==`, `!=`, `>`, `<`, `>=`, `<=`, `&&`, `||`, `!`, `+`, `-`, `*`, `/`, `%`
- member and index access
- function calls and method calls
- list literals and regex literals
- `this` binding with embed-aware scoping
- list helpers (`filter`, `map`, `reduce`) using implicit locals `value`, `index`, `acc`
- summary context helpers: `values` (column values) and `rows` (row files)
-
-## diagnostics
-
- parser diagnostics are collected with spans at compile time
- runtime diagnostics are collected during evaluation and deduped per context
- base views render diagnostics above the view output
-
-## this scoping
-
- main base file: `this` resolves to the base file
- embedded base: `this` resolves to the embedding file
- row evaluation: `file` resolves to the row file
-
-## performance decisions
-
- bytecode ir keeps evaluation linear and stable
- per-build backlink index avoids n^2 scans
- property cache memoizes property expressions per file
- formula cache memoizes formula evaluation per file
-
-## view rendering
-
- table, list, cards/gallery, board, calendar, map
- map rendering expects coordinates `[lat, lon]` and map config fields
- view filters combine with base filters via logical and
--- a/quartz/util/base/compiler/ast.ts
+++ b/quartz/util/base/compiler/ast.ts
@@ -1,76 +0,0 @@
-export type Position = { offset: number; line: number; column: number }
-
-export type Span = { start: Position; end: Position; file?: string }
-
-export type Program = { type: "Program"; body: Expr | null; span: Span }
-
-export type Expr =
-  | Literal
-  | Identifier
-  | UnaryExpr
-  | BinaryExpr
-  | LogicalExpr
-  | CallExpr
-  | MemberExpr
-  | IndexExpr
-  | ListExpr
-  | ErrorExpr
-
-export type LiteralKind = "number" | "string" | "boolean" | "null" | "date" | "duration" | "regex"
-
-export type NumberLiteral = { type: "Literal"; kind: "number"; value: number; span: Span }
-export type StringLiteral = { type: "Literal"; kind: "string"; value: string; span: Span }
-export type BooleanLiteral = { type: "Literal"; kind: "boolean"; value: boolean; span: Span }
-export type NullLiteral = { type: "Literal"; kind: "null"; value: null; span: Span }
-export type DateLiteral = { type: "Literal"; kind: "date"; value: string; span: Span }
-export type DurationLiteral = { type: "Literal"; kind: "duration"; value: string; span: Span }
-export type RegexLiteral = {
-  type: "Literal"
-  kind: "regex"
-  value: string
-  flags: string
-  span: Span
-}
-
-export type Literal =
-  | NumberLiteral
-  | StringLiteral
-  | BooleanLiteral
-  | NullLiteral
-  | DateLiteral
-  | DurationLiteral
-  | RegexLiteral
-
-export type Identifier = { type: "Identifier"; name: string; span: Span }
-
-export type UnaryExpr = { type: "UnaryExpr"; operator: "!" | "-"; argument: Expr; span: Span }
-
-export type BinaryExpr = {
-  type: "BinaryExpr"
-  operator: "+" | "-" | "*" | "/" | "%" | "==" | "!=" | ">" | ">=" | "<" | "<="
-  left: Expr
-  right: Expr
-  span: Span
-}
-
-export type LogicalExpr = {
-  type: "LogicalExpr"
-  operator: "&&" | "||"
-  left: Expr
-  right: Expr
-  span: Span
-}
-
-export type CallExpr = { type: "CallExpr"; callee: Expr; args: Expr[]; span: Span }
-
-export type MemberExpr = { type: "MemberExpr"; object: Expr; property: string; span: Span }
-
-export type IndexExpr = { type: "IndexExpr"; object: Expr; index: Expr; span: Span }
-
-export type ListExpr = { type: "ListExpr"; elements: Expr[]; span: Span }
-
-export type ErrorExpr = { type: "ErrorExpr"; message: string; span: Span }
-
-export function spanFrom(start: Span, end: Span): Span {
-  return { start: start.start, end: end.end, file: start.file || end.file }
-}
--- a/quartz/util/base/compiler/diagnostics.ts
+++ b/quartz/util/base/compiler/diagnostics.ts
@@ -1,9 +0,0 @@
-import { Span } from "./ast"
-
-export type BaseExpressionDiagnostic = {
-  kind: "lex" | "parse" | "runtime"
-  message: string
-  span: Span
-  context: string
-  source: string
-}
--- a/quartz/util/base/compiler/errors.ts
+++ b/quartz/util/base/compiler/errors.ts
@@ -1,3 +0,0 @@
-import { Span } from "./ast"
-
-export type Diagnostic = { kind: "lex" | "parse"; message: string; span: Span }
--- a/quartz/util/base/compiler/expressions.ts
+++ b/quartz/util/base/compiler/expressions.ts
@@ -1,10 +0,0 @@
-import { ProgramIR } from "./ir"
-
-export type BasesExpressions = {
-  filters?: ProgramIR
-  viewFilters: Record<string, ProgramIR>
-  formulas: Record<string, ProgramIR>
-  summaries: Record<string, ProgramIR>
-  viewSummaries: Record<string, Record<string, ProgramIR>>
-  propertyExpressions: Record<string, ProgramIR>
-}
--- a/quartz/util/base/compiler/index.ts
+++ b/quartz/util/base/compiler/index.ts
@@ -1,44 +0,0 @@
-export { lex } from "./lexer"
-export { parseExpressionSource } from "./parser"
-export type { ParseResult } from "./parser"
-export type { Diagnostic } from "./errors"
-export type { Program, Expr, Span, Position } from "./ast"
-export type { BaseExpressionDiagnostic } from "./diagnostics"
-export type { BasesExpressions } from "./expressions"
-export type { Instruction, ProgramIR } from "./ir"
-export { compileExpression } from "./ir"
-export { buildPropertyExpressionSource } from "./properties"
-export type {
-  SummaryDefinition,
-  ViewSummaryConfig,
-  PropertyConfig,
-  BuiltinSummaryType,
-} from "./schema"
-export { BUILTIN_SUMMARY_TYPES } from "./schema"
-export {
-  evaluateExpression,
-  evaluateFilterExpression,
-  evaluateSummaryExpression,
-  valueToUnknown,
-} from "./interpreter"
-export type {
-  EvalContext,
-  Value,
-  NullValue,
-  BooleanValue,
-  NumberValue,
-  StringValue,
-  DateValue,
-  DurationValue,
-  ListValue,
-  ObjectValue,
-  FileValue,
-  LinkValue,
-  RegexValue,
-  HtmlValue,
-  IconValue,
-  ImageValue,
-  ValueKind,
-  ValueOf,
-} from "./interpreter"
-export { isValueKind } from "./interpreter"
--- a/quartz/util/base/compiler/interpreter.test.ts
+++ b/quartz/util/base/compiler/interpreter.test.ts
@@ -1,73 +0,0 @@
-import assert from "node:assert"
-import test from "node:test"
-import { FilePath, FullSlug, SimpleSlug } from "../../path"
-
-type ContentLayout = "default" | "article" | "page"
-import { evaluateExpression, valueToUnknown, EvalContext } from "./interpreter"
-import { compileExpression } from "./ir"
-import { parseExpressionSource } from "./parser"
-
-const parseExpr = (source: string) => {
-  const result = parseExpressionSource(source, "test")
-  if (!result.program.body) {
-    throw new Error(`expected expression for ${source}`)
-  }
-  return compileExpression(result.program.body)
-}
-
-const makeCtx = (): EvalContext => {
-  const fileA = {
-    slug: "a" as FullSlug,
-    filePath: "a.md" as FilePath,
-    frontmatter: { title: "A", pageLayout: "default" as ContentLayout },
-    links: [] as SimpleSlug[],
-  }
-  const fileB = {
-    slug: "b" as FullSlug,
-    filePath: "b.md" as FilePath,
-    frontmatter: { title: "B", pageLayout: "default" as ContentLayout },
-    links: ["a"] as SimpleSlug[],
-  }
-  return { file: fileA, allFiles: [fileA, fileB] }
-}
-
-test("link equality resolves to file targets", () => {
-  const expr = parseExpr('link("a") == file("a")')
-  const value = valueToUnknown(evaluateExpression(expr, makeCtx()))
-  assert.strictEqual(value, true)
-})
-
-test("link equality matches raw string targets", () => {
-  const expr = parseExpr('link("a") == "a"')
-  const value = valueToUnknown(evaluateExpression(expr, makeCtx()))
-  assert.strictEqual(value, true)
-})
-
-test("date arithmetic handles month additions", () => {
-  const expr = parseExpr('date("2025-01-01") + "1M"')
-  const value = valueToUnknown(evaluateExpression(expr, makeCtx()))
-  assert.ok(value instanceof Date)
-  assert.strictEqual(value.toISOString().split("T")[0], "2025-02-01")
-})
-
-test("date subtraction returns duration in ms", () => {
-  const expr = parseExpr('date("2025-01-02") - date("2025-01-01")')
-  const value = valueToUnknown(evaluateExpression(expr, makeCtx()))
-  assert.strictEqual(value, 86400000)
-})
-
-test("list summary helpers compute statistics", () => {
-  const meanExpr = parseExpr("([1, 2, 3]).mean()")
-  const medianExpr = parseExpr("([1, 2, 3]).median()")
-  const stddevExpr = parseExpr("([1, 2, 3]).stddev()")
-  const sumExpr = parseExpr("([1, 2, 3]).sum()")
-  const ctx = makeCtx()
-  assert.strictEqual(valueToUnknown(evaluateExpression(meanExpr, ctx)), 2)
-  assert.strictEqual(valueToUnknown(evaluateExpression(medianExpr, ctx)), 2)
-  assert.strictEqual(valueToUnknown(evaluateExpression(sumExpr, ctx)), 6)
-  const stddev = valueToUnknown(evaluateExpression(stddevExpr, ctx))
-  assert.strictEqual(typeof stddev, "number")
-  if (typeof stddev === "number") {
-    assert.ok(Math.abs(stddev - Math.sqrt(2 / 3)) < 1e-6)
-  }
-})
--- a/quartz/util/base/compiler/interpreter.ts
+++ b/quartz/util/base/compiler/interpreter.ts
--- a/quartz/util/base/compiler/ir.ts
+++ b/quartz/util/base/compiler/ir.ts
@@ -1,164 +0,0 @@
-import { BinaryExpr, Expr, Literal, Span, UnaryExpr } from "./ast"
-
-export type JumpInstruction = {
-  op: "jump" | "jump_if_false" | "jump_if_true"
-  target: number
-  span: Span
-}
-
-export type Instruction =
-  | { op: "const"; literal: Literal; span: Span }
-  | { op: "ident"; name: string; span: Span }
-  | { op: "load_formula"; name: string; span: Span }
-  | { op: "load_formula_index"; span: Span }
-  | { op: "member"; property: string; span: Span }
-  | { op: "index"; span: Span }
-  | { op: "list"; count: number; span: Span }
-  | { op: "unary"; operator: UnaryExpr["operator"]; span: Span }
-  | { op: "binary"; operator: BinaryExpr["operator"]; span: Span }
-  | { op: "to_bool"; span: Span }
-  | { op: "call_global"; name: string; argc: number; span: Span }
-  | { op: "call_method"; name: string; argc: number; span: Span }
-  | { op: "call_dynamic"; span: Span }
-  | { op: "filter"; program: ProgramIR | null; span: Span }
-  | { op: "map"; program: ProgramIR | null; span: Span }
-  | { op: "reduce"; program: ProgramIR | null; initial: ProgramIR | null; span: Span }
-  | JumpInstruction
-
-export type ProgramIR = { instructions: Instruction[]; span: Span }
-
-const compileExpr = (expr: Expr, out: Instruction[]) => {
-  switch (expr.type) {
-    case "Literal":
-      out.push({ op: "const", literal: expr, span: expr.span })
-      return
-    case "Identifier":
-      out.push({ op: "ident", name: expr.name, span: expr.span })
-      return
-    case "UnaryExpr":
-      compileExpr(expr.argument, out)
-      out.push({ op: "unary", operator: expr.operator, span: expr.span })
-      return
-    case "BinaryExpr":
-      compileExpr(expr.left, out)
-      compileExpr(expr.right, out)
-      out.push({ op: "binary", operator: expr.operator, span: expr.span })
-      return
-    case "LogicalExpr": {
-      if (expr.operator === "&&") {
-        compileExpr(expr.left, out)
-        const jumpFalse: JumpInstruction = { op: "jump_if_false", target: -1, span: expr.span }
-        out.push(jumpFalse)
-        compileExpr(expr.right, out)
-        out.push({ op: "to_bool", span: expr.span })
-        const jumpEnd: JumpInstruction = { op: "jump", target: -1, span: expr.span }
-        out.push(jumpEnd)
-        const falseTarget = out.length
-        jumpFalse.target = falseTarget
-        out.push({
-          op: "const",
-          literal: { type: "Literal", kind: "boolean", value: false, span: expr.span },
-          span: expr.span,
-        })
-        jumpEnd.target = out.length
-        return
-      }
-      compileExpr(expr.left, out)
-      const jumpTrue: JumpInstruction = { op: "jump_if_true", target: -1, span: expr.span }
-      out.push(jumpTrue)
-      compileExpr(expr.right, out)
-      out.push({ op: "to_bool", span: expr.span })
-      const jumpEnd: JumpInstruction = { op: "jump", target: -1, span: expr.span }
-      out.push(jumpEnd)
-      const trueTarget = out.length
-      jumpTrue.target = trueTarget
-      out.push({
-        op: "const",
-        literal: { type: "Literal", kind: "boolean", value: true, span: expr.span },
-        span: expr.span,
-      })
-      jumpEnd.target = out.length
-      return
-    }
-    case "MemberExpr":
-      if (expr.object.type === "Identifier" && expr.object.name === "formula") {
-        out.push({ op: "load_formula", name: expr.property, span: expr.span })
-        return
-      }
-      compileExpr(expr.object, out)
-      out.push({ op: "member", property: expr.property, span: expr.span })
-      return
-    case "IndexExpr":
-      if (expr.object.type === "Identifier" && expr.object.name === "formula") {
-        compileExpr(expr.index, out)
-        out.push({ op: "load_formula_index", span: expr.span })
-        return
-      }
-      compileExpr(expr.object, out)
-      compileExpr(expr.index, out)
-      out.push({ op: "index", span: expr.span })
-      return
-    case "ListExpr":
-      for (const element of expr.elements) {
-        compileExpr(element, out)
-      }
-      out.push({ op: "list", count: expr.elements.length, span: expr.span })
-      return
-    case "CallExpr": {
-      if (expr.callee.type === "Identifier") {
-        for (const arg of expr.args) {
-          compileExpr(arg, out)
-        }
-        out.push({
-          op: "call_global",
-          name: expr.callee.name,
-          argc: expr.args.length,
-          span: expr.span,
-        })
-        return
-      }
-      if (expr.callee.type === "MemberExpr") {
-        const method = expr.callee.property
-        if (method === "filter" || method === "map" || method === "reduce") {
-          compileExpr(expr.callee.object, out)
-          const exprArg = expr.args[0]
-          const program = exprArg ? compileExpression(exprArg) : null
-          if (method === "filter") {
-            out.push({ op: "filter", program, span: expr.span })
-            return
-          }
-          if (method === "map") {
-            out.push({ op: "map", program, span: expr.span })
-            return
-          }
-          const initialArg = expr.args[1]
-          const initial = initialArg ? compileExpression(initialArg) : null
-          out.push({ op: "reduce", program, initial, span: expr.span })
-          return
-        }
-        compileExpr(expr.callee.object, out)
-        for (const arg of expr.args) {
-          compileExpr(arg, out)
-        }
-        out.push({ op: "call_method", name: method, argc: expr.args.length, span: expr.span })
-        return
-      }
-      compileExpr(expr.callee, out)
-      out.push({ op: "call_dynamic", span: expr.span })
-      return
-    }
-    case "ErrorExpr":
-      out.push({
-        op: "const",
-        literal: { type: "Literal", kind: "null", value: null, span: expr.span },
-        span: expr.span,
-      })
-      return
-  }
-}
-
-export const compileExpression = (expr: Expr): ProgramIR => {
-  const instructions: Instruction[] = []
-  compileExpr(expr, instructions)
-  return { instructions, span: expr.span }
-}
--- a/quartz/util/base/compiler/lexer.test.ts
+++ b/quartz/util/base/compiler/lexer.test.ts
@@ -1,53 +0,0 @@
-import assert from "node:assert"
-import test from "node:test"
-import { lex } from "./lexer"
-
-test("lexes bracket access with hyphenated keys", () => {
-  const result = lex('note["my-field"]')
-  const types = result.tokens.map((token) => token.type)
-  assert.deepStrictEqual(types, ["identifier", "punctuation", "string", "punctuation", "eof"])
-  const value = result.tokens[2]
-  if (value.type !== "string") {
-    throw new Error("expected string token")
-  }
-  assert.strictEqual(value.value, "my-field")
-})
-
-test("lexes bracket access with escaped quotes", () => {
-  const result = lex('note["my\\\"field"]')
-  const value = result.tokens.find((token) => token.type === "string")
-  if (!value || value.type !== "string") {
-    throw new Error("expected string token")
-  }
-  assert.strictEqual(value.value, 'my"field')
-})
-
-test("lexes regex literals with flags", () => {
-  const result = lex('name.replace(/:/g, "-")')
-  const regexToken = result.tokens.find((token) => token.type === "regex")
-  if (!regexToken || regexToken.type !== "regex") {
-    throw new Error("expected regex token")
-  }
-  assert.strictEqual(regexToken.pattern, ":")
-  assert.strictEqual(regexToken.flags, "g")
-})
-
-test("lexes regex literals with escaped slashes", () => {
-  const result = lex("path.matches(/\\//)")
-  const regexToken = result.tokens.find((token) => token.type === "regex")
-  if (!regexToken || regexToken.type !== "regex") {
-    throw new Error("expected regex token")
-  }
-  assert.strictEqual(regexToken.pattern, "\\/")
-  assert.strictEqual(regexToken.flags, "")
-})
-
-test("lexes division as operator, not regex", () => {
-  const result = lex("a / b")
-  const operatorToken = result.tokens.find(
-    (token) => token.type === "operator" && token.value === "/",
-  )
-  assert.ok(operatorToken)
-  const regexToken = result.tokens.find((token) => token.type === "regex")
-  assert.strictEqual(regexToken, undefined)
-})
--- a/quartz/util/base/compiler/lexer.ts
+++ b/quartz/util/base/compiler/lexer.ts
@@ -1,300 +0,0 @@
-import { Position, Span } from "./ast"
-import { Diagnostic } from "./errors"
-import {
-  Operator,
-  Punctuation,
-  Token,
-  StringToken,
-  RegexToken,
-  NumberToken,
-  BooleanToken,
-  NullToken,
-  ThisToken,
-  IdentifierToken,
-  OperatorToken,
-  PunctuationToken,
-  EofToken,
-} from "./tokens"
-
-type LexResult = { tokens: Token[]; diagnostics: Diagnostic[] }
-
-const operatorTokens: Operator[] = [
-  "==",
-  "!=",
-  ">=",
-  "<=",
-  "&&",
-  "||",
-  "+",
-  "-",
-  "*",
-  "/",
-  "%",
-  "!",
-  ">",
-  "<",
-]
-
-const punctuationTokens: Punctuation[] = [".", ",", "(", ")", "[", "]"]
-
-const isOperator = (value: string): value is Operator =>
-  operatorTokens.some((token) => token === value)
-
-const isPunctuation = (value: string): value is Punctuation =>
-  punctuationTokens.some((token) => token === value)
-
-export function lex(input: string, file?: string): LexResult {
-  const tokens: Token[] = []
-  const diagnostics: Diagnostic[] = []
-  let index = 0
-  let line = 1
-  let column = 1
-  let canStartRegex = true
-
-  const makePosition = (offset: number, lineValue: number, columnValue: number): Position => ({
-    offset,
-    line: lineValue,
-    column: columnValue,
-  })
-
-  const currentPosition = (): Position => makePosition(index, line, column)
-
-  const makeSpan = (start: Position, end: Position): Span => ({ start, end, file })
-
-  const advance = (): string => {
-    const ch = input[index]
-    index += 1
-    if (ch === "\n") {
-      line += 1
-      column = 1
-    } else {
-      column += 1
-    }
-    return ch
-  }
-
-  const peek = (offset = 0): string => input[index + offset] ?? ""
-
-  const addDiagnostic = (message: string, span: Span) => {
-    diagnostics.push({ kind: "lex", message, span })
-  }
-
-  const updateRegexState = (token: Token | null) => {
-    if (!token) {
-      canStartRegex = true
-      return
-    }
-    if (token.type === "operator") {
-      canStartRegex = true
-      return
-    }
-    if (token.type === "punctuation") {
-      canStartRegex = token.value === "(" || token.value === "[" || token.value === ","
-      return
-    }
-    canStartRegex = false
-  }
-
-  const isWhitespace = (ch: string) => ch === " " || ch === "\t" || ch === "\n" || ch === "\r"
-  const isDigit = (ch: string) => ch >= "0" && ch <= "9"
-  const isIdentStart = (ch: string) =>
-    (ch >= "a" && ch <= "z") || (ch >= "A" && ch <= "Z") || ch === "_"
-  const isIdentContinue = (ch: string) => isIdentStart(ch) || isDigit(ch)
-
-  while (index < input.length) {
-    const ch = peek()
-
-    if (isWhitespace(ch)) {
-      advance()
-      continue
-    }
-
-    const start = currentPosition()
-
-    if (ch === "=" && peek(1) !== "=") {
-      let offset = 1
-      while (isWhitespace(peek(offset))) {
-        offset += 1
-      }
-      if (peek(offset) === ">") {
-        advance()
-        for (let step = 1; step < offset; step += 1) {
-          advance()
-        }
-        if (peek() === ">") {
-          advance()
-        }
-        const end = currentPosition()
-        addDiagnostic(
-          "arrow functions are not supported, use list.filter(expression)",
-          makeSpan(start, end),
-        )
-        continue
-      }
-    }
-
-    if (ch === '"' || ch === "'") {
-      const quote = advance()
-      let value = ""
-      let closed = false
-
-      while (index < input.length) {
-        const curr = advance()
-        if (curr === quote) {
-          closed = true
-          break
-        }
-        if (curr === "\\") {
-          const next = advance()
-          if (next === "n") value += "\n"
-          else if (next === "t") value += "\t"
-          else if (next === "r") value += "\r"
-          else if (next === "\\" || next === "'" || next === '"') value += next
-          else value += next
-        } else {
-          value += curr
-        }
-      }
-
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      if (!closed) addDiagnostic("unterminated string literal", span)
-      const token: StringToken = { type: "string", value, span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    if (ch === "/" && canStartRegex) {
-      const next = peek(1)
-      if (next !== "/" && next !== "") {
-        advance()
-        let pattern = ""
-        let closed = false
-        let inClass = false
-        while (index < input.length) {
-          const curr = advance()
-          if (curr === "\\" && index < input.length) {
-            const escaped = advance()
-            pattern += `\\${escaped}`
-            continue
-          }
-          if (curr === "[" && !inClass) inClass = true
-          if (curr === "]" && inClass) inClass = false
-          if (curr === "/" && !inClass) {
-            closed = true
-            break
-          }
-          pattern += curr
-        }
-        let flags = ""
-        while (index < input.length) {
-          const flag = peek()
-          if (!/^[gimsuy]$/.test(flag)) break
-          flags += advance()
-        }
-        const end = currentPosition()
-        const span = makeSpan(start, end)
-        if (!closed) addDiagnostic("unterminated regex literal", span)
-        const token: RegexToken = { type: "regex", pattern, flags, span }
-        tokens.push(token)
-        updateRegexState(token)
-        continue
-      }
-    }
-
-    if (isDigit(ch)) {
-      let num = ""
-      while (index < input.length && isDigit(peek())) {
-        num += advance()
-      }
-      if (peek() === "." && isDigit(peek(1))) {
-        num += advance()
-        while (index < input.length && isDigit(peek())) {
-          num += advance()
-        }
-      }
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      const token: NumberToken = { type: "number", value: Number(num), span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    if (isIdentStart(ch)) {
-      let ident = ""
-      while (index < input.length && isIdentContinue(peek())) {
-        ident += advance()
-      }
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      if (ident === "true" || ident === "false") {
-        const token: BooleanToken = { type: "boolean", value: ident === "true", span }
-        tokens.push(token)
-        updateRegexState(token)
-        continue
-      }
-      if (ident === "null") {
-        const token: NullToken = { type: "null", span }
-        tokens.push(token)
-        updateRegexState(token)
-        continue
-      }
-      if (ident === "this") {
-        const token: ThisToken = { type: "this", span }
-        tokens.push(token)
-        updateRegexState(token)
-        continue
-      }
-      const token: IdentifierToken = { type: "identifier", value: ident, span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    const twoChar = ch + peek(1)
-    if (isOperator(twoChar)) {
-      advance()
-      advance()
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      const token: OperatorToken = { type: "operator", value: twoChar, span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    if (isOperator(ch)) {
-      advance()
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      const token: OperatorToken = { type: "operator", value: ch, span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    if (isPunctuation(ch)) {
-      advance()
-      const end = currentPosition()
-      const span = makeSpan(start, end)
-      const token: PunctuationToken = { type: "punctuation", value: ch, span }
-      tokens.push(token)
-      updateRegexState(token)
-      continue
-    }
-
-    advance()
-    const end = currentPosition()
-    addDiagnostic(`unexpected character: ${ch}`, makeSpan(start, end))
-  }
-
-  const eofPos = currentPosition()
-  const eofSpan = makeSpan(eofPos, eofPos)
-  const eofToken: EofToken = { type: "eof", span: eofSpan }
-  tokens.push(eofToken)
-  updateRegexState(eofToken)
-
-  return { tokens, diagnostics }
-}
--- a/quartz/util/base/compiler/parser.test.ts
+++ b/quartz/util/base/compiler/parser.test.ts
@@ -1,261 +0,0 @@
-import assert from "node:assert"
-import test from "node:test"
-import { parseExpressionSource } from "./parser"
-
-const isRecord = (value: unknown): value is Record<string, unknown> =>
-  typeof value === "object" && value !== null
-
-const strip = (node: unknown): unknown => {
-  if (!isRecord(node)) return node
-  const type = node.type
-  if (type === "Identifier") {
-    return { type, name: node.name }
-  }
-  if (type === "Literal") {
-    const kind = node.kind
-    const value = node.value
-    const flags = node.flags
-    return flags !== undefined ? { type, kind, value, flags } : { type, kind, value }
-  }
-  if (type === "UnaryExpr") {
-    return { type, operator: node.operator, argument: strip(node.argument) }
-  }
-  if (type === "BinaryExpr" || type === "LogicalExpr") {
-    return { type, operator: node.operator, left: strip(node.left), right: strip(node.right) }
-  }
-  if (type === "CallExpr") {
-    const args = Array.isArray(node.args) ? node.args.map(strip) : []
-    return { type, callee: strip(node.callee), args }
-  }
-  if (type === "MemberExpr") {
-    return { type, object: strip(node.object), property: node.property }
-  }
-  if (type === "IndexExpr") {
-    return { type, object: strip(node.object), index: strip(node.index) }
-  }
-  if (type === "ListExpr") {
-    const elements = Array.isArray(node.elements) ? node.elements.map(strip) : []
-    return { type, elements }
-  }
-  if (type === "ErrorExpr") {
-    return { type, message: node.message }
-  }
-  return node
-}
-
-test("ebnf to ast mapping snapshots", () => {
-  const cases: Array<{ source: string; expected: unknown }> = [
-    {
-      source: 'status == "done"',
-      expected: {
-        type: "BinaryExpr",
-        operator: "==",
-        left: { type: "Identifier", name: "status" },
-        right: { type: "Literal", kind: "string", value: "done" },
-      },
-    },
-    {
-      source: "!done",
-      expected: {
-        type: "UnaryExpr",
-        operator: "!",
-        argument: { type: "Identifier", name: "done" },
-      },
-    },
-    {
-      source: "file.ctime",
-      expected: {
-        type: "MemberExpr",
-        object: { type: "Identifier", name: "file" },
-        property: "ctime",
-      },
-    },
-    {
-      source: 'note["my-field"]',
-      expected: {
-        type: "IndexExpr",
-        object: { type: "Identifier", name: "note" },
-        index: { type: "Literal", kind: "string", value: "my-field" },
-      },
-    },
-    {
-      source: "date(due) < today()",
-      expected: {
-        type: "BinaryExpr",
-        operator: "<",
-        left: {
-          type: "CallExpr",
-          callee: { type: "Identifier", name: "date" },
-          args: [{ type: "Identifier", name: "due" }],
-        },
-        right: { type: "CallExpr", callee: { type: "Identifier", name: "today" }, args: [] },
-      },
-    },
-    {
-      source: "now() - file.ctime",
-      expected: {
-        type: "BinaryExpr",
-        operator: "-",
-        left: { type: "CallExpr", callee: { type: "Identifier", name: "now" }, args: [] },
-        right: {
-          type: "MemberExpr",
-          object: { type: "Identifier", name: "file" },
-          property: "ctime",
-        },
-      },
-    },
-    {
-      source: "(pages * 2).round(0)",
-      expected: {
-        type: "CallExpr",
-        callee: {
-          type: "MemberExpr",
-          object: {
-            type: "BinaryExpr",
-            operator: "*",
-            left: { type: "Identifier", name: "pages" },
-            right: { type: "Literal", kind: "number", value: 2 },
-          },
-          property: "round",
-        },
-        args: [{ type: "Literal", kind: "number", value: 0 }],
-      },
-    },
-    {
-      source: 'tags.containsAny("a","b")',
-      expected: {
-        type: "CallExpr",
-        callee: {
-          type: "MemberExpr",
-          object: { type: "Identifier", name: "tags" },
-          property: "containsAny",
-        },
-        args: [
-          { type: "Literal", kind: "string", value: "a" },
-          { type: "Literal", kind: "string", value: "b" },
-        ],
-      },
-    },
-    {
-      source: "list(links).filter(value.isTruthy())",
-      expected: {
-        type: "CallExpr",
-        callee: {
-          type: "MemberExpr",
-          object: {
-            type: "CallExpr",
-            callee: { type: "Identifier", name: "list" },
-            args: [{ type: "Identifier", name: "links" }],
-          },
-          property: "filter",
-        },
-        args: [
-          {
-            type: "CallExpr",
-            callee: {
-              type: "MemberExpr",
-              object: { type: "Identifier", name: "value" },
-              property: "isTruthy",
-            },
-            args: [],
-          },
-        ],
-      },
-    },
-    {
-      source: '["a", "b", "c"].length',
-      expected: {
-        type: "MemberExpr",
-        object: {
-          type: "ListExpr",
-          elements: [
-            { type: "Literal", kind: "string", value: "a" },
-            { type: "Literal", kind: "string", value: "b" },
-            { type: "Literal", kind: "string", value: "c" },
-          ],
-        },
-        property: "length",
-      },
-    },
-    {
-      source: "this.file.name",
-      expected: {
-        type: "MemberExpr",
-        object: {
-          type: "MemberExpr",
-          object: { type: "Identifier", name: "this" },
-          property: "file",
-        },
-        property: "name",
-      },
-    },
-    {
-      source: "a || b && c",
-      expected: {
-        type: "LogicalExpr",
-        operator: "||",
-        left: { type: "Identifier", name: "a" },
-        right: {
-          type: "LogicalExpr",
-          operator: "&&",
-          left: { type: "Identifier", name: "b" },
-          right: { type: "Identifier", name: "c" },
-        },
-      },
-    },
-    {
-      source: "values[0]",
-      expected: {
-        type: "IndexExpr",
-        object: { type: "Identifier", name: "values" },
-        index: { type: "Literal", kind: "number", value: 0 },
-      },
-    },
-  ]
-
-  for (const entry of cases) {
-    const result = parseExpressionSource(entry.source)
-    assert.strictEqual(result.diagnostics.length, 0)
-    assert.deepStrictEqual(strip(result.program.body), entry.expected)
-  }
-})
-
-test("syntax doc samples parse", () => {
-  const samples = [
-    'note["price"]',
-    "file.size > 10",
-    "file.hasLink(this.file)",
-    'date("2024-12-01") + "1M" + "4h" + "3m"',
-    "now() - file.ctime",
-    "property[0]",
-    'link("filename", icon("plus"))',
-    'file.mtime > now() - "1 week"',
-    '/abc/.matches("abcde")',
-    'name.replace(/:/g, "-")',
-    'values.filter(value.isType("number")).reduce(if(acc == null || value > acc, value, acc), null)',
-  ]
-
-  for (const source of samples) {
-    const result = parseExpressionSource(source)
-    assert.strictEqual(result.diagnostics.length, 0)
-    assert.ok(result.program.body)
-  }
-})
-
-test("string escapes are decoded", () => {
-  const result = parseExpressionSource('"a\\n\\"b"')
-  assert.strictEqual(result.diagnostics.length, 0)
-  const literal = strip(result.program.body)
-  if (!isRecord(literal)) {
-    throw new Error("expected literal record")
-  }
-  assert.strictEqual(literal.type, "Literal")
-  assert.strictEqual(literal.kind, "string")
-  assert.strictEqual(literal.value, 'a\n"b')
-})
-
-test("parser reports errors and recovers", () => {
-  const result = parseExpressionSource("status ==")
-  assert.ok(result.diagnostics.length > 0)
-  assert.ok(result.program.body)
-})
--- a/quartz/util/base/compiler/parser.ts
+++ b/quartz/util/base/compiler/parser.ts
@@ -1,370 +0,0 @@
-import {
-  BinaryExpr,
-  CallExpr,
-  ErrorExpr,
-  Expr,
-  Identifier,
-  IndexExpr,
-  ListExpr,
-  Literal,
-  LogicalExpr,
-  MemberExpr,
-  Program,
-  UnaryExpr,
-  spanFrom,
-} from "./ast"
-import { Diagnostic } from "./errors"
-import { lex } from "./lexer"
-import { Operator, Token } from "./tokens"
-
-export type ParseResult = { program: Program; tokens: Token[]; diagnostics: Diagnostic[] }
-
-type InfixInfo = { lbp: number; rbp: number; kind: "binary" | "logical" }
-
-const infixBindingPowers: Record<string, InfixInfo> = {
-  "||": { lbp: 1, rbp: 2, kind: "logical" },
-  "&&": { lbp: 3, rbp: 4, kind: "logical" },
-  "==": { lbp: 5, rbp: 6, kind: "binary" },
-  "!=": { lbp: 5, rbp: 6, kind: "binary" },
-  ">": { lbp: 7, rbp: 8, kind: "binary" },
-  ">=": { lbp: 7, rbp: 8, kind: "binary" },
-  "<": { lbp: 7, rbp: 8, kind: "binary" },
-  "<=": { lbp: 7, rbp: 8, kind: "binary" },
-  "+": { lbp: 9, rbp: 10, kind: "binary" },
-  "-": { lbp: 9, rbp: 10, kind: "binary" },
-  "*": { lbp: 11, rbp: 12, kind: "binary" },
-  "/": { lbp: 11, rbp: 12, kind: "binary" },
-  "%": { lbp: 11, rbp: 12, kind: "binary" },
-}
-
-const isLogicalOperator = (value: Operator): value is LogicalExpr["operator"] =>
-  value === "&&" || value === "||"
-
-const isBinaryOperator = (value: Operator): value is BinaryExpr["operator"] =>
-  value === "+" ||
-  value === "-" ||
-  value === "*" ||
-  value === "/" ||
-  value === "%" ||
-  value === "==" ||
-  value === "!=" ||
-  value === ">" ||
-  value === ">=" ||
-  value === "<" ||
-  value === "<="
-
-export function parseExpressionSource(source: string, file?: string): ParseResult {
-  const { tokens, diagnostics } = lex(source, file)
-  const parser = new Parser(tokens, diagnostics)
-  const program = parser.parseProgram()
-  return { program, tokens, diagnostics }
-}
-
-class Parser {
-  private tokens: Token[]
-  private index: number
-  private diagnostics: Diagnostic[]
-
-  constructor(tokens: Token[], diagnostics: Diagnostic[]) {
-    this.tokens = tokens
-    this.index = 0
-    this.diagnostics = diagnostics
-  }
-
-  parseProgram(): Program {
-    const start = this.tokens[0]?.span ?? this.tokens[this.tokens.length - 1].span
-    const body = this.peek().type === "eof" ? null : this.parseExpression(0)
-    const end = this.tokens[this.tokens.length - 1]?.span ?? start
-    return { type: "Program", body, span: spanFrom(start, end) }
-  }
-
-  private parseExpression(minBp: number): Expr {
-    let left = this.parsePrefix()
-    left = this.parsePostfix(left)
-
-    while (true) {
-      const token = this.peek()
-      if (token.type !== "operator") break
-      const info = infixBindingPowers[token.value]
-      if (!info || info.lbp < minBp) break
-      this.advance()
-      const right = this.parseExpression(info.rbp)
-      const span = spanFrom(left.span, right.span)
-      if (info.kind === "logical" && isLogicalOperator(token.value)) {
-        left = { type: "LogicalExpr", operator: token.value, left, right, span }
-      } else if (info.kind === "binary" && isBinaryOperator(token.value)) {
-        left = { type: "BinaryExpr", operator: token.value, left, right, span }
-      } else {
-        this.error("unexpected operator", token.span)
-      }
-    }
-
-    return left
-  }
-
-  private parsePrefix(): Expr {
-    const token = this.peek()
-    if (token.type === "operator" && (token.value === "!" || token.value === "-")) {
-      this.advance()
-      const argument = this.parseExpression(13)
-      const span = spanFrom(token.span, argument.span)
-      const node: UnaryExpr = { type: "UnaryExpr", operator: token.value, argument, span }
-      return node
-    }
-    return this.parsePrimary()
-  }
-
-  private parsePostfix(expr: Expr): Expr {
-    let current = expr
-    while (true) {
-      const token = this.peek()
-      if (token.type === "punctuation" && token.value === ".") {
-        this.advance()
-        const propToken = this.peek()
-        if (propToken.type !== "identifier") {
-          this.error("expected identifier after '.'", propToken.span)
-          return current
-        }
-        this.advance()
-        const span = spanFrom(current.span, propToken.span)
-        const node: MemberExpr = {
-          type: "MemberExpr",
-          object: current,
-          property: propToken.value,
-          span,
-        }
-        current = node
-        continue
-      }
-
-      if (token.type === "punctuation" && token.value === "[") {
-        this.advance()
-        const indexExpr = this.parseExpression(0)
-        const endToken = this.peek()
-        if (!(endToken.type === "punctuation" && endToken.value === "]")) {
-          this.error("expected ']'", endToken.span)
-          this.syncTo("]")
-        } else {
-          this.advance()
-        }
-        const span = spanFrom(current.span, endToken.span)
-        const node: IndexExpr = { type: "IndexExpr", object: current, index: indexExpr, span }
-        current = node
-        continue
-      }
-
-      if (token.type === "punctuation" && token.value === "(") {
-        this.advance()
-        const args: Expr[] = []
-        while (this.peek().type !== "eof") {
-          const next = this.peek()
-          if (next.type === "punctuation" && next.value === ")") {
-            this.advance()
-            break
-          }
-          const arg = this.parseExpression(0)
-          args.push(arg)
-          const sep = this.peek()
-          if (sep.type === "punctuation" && sep.value === ",") {
-            this.advance()
-            const maybeClose = this.peek()
-            if (maybeClose.type === "punctuation" && maybeClose.value === ")") {
-              this.advance()
-              break
-            }
-            continue
-          }
-          if (sep.type === "punctuation" && sep.value === ")") {
-            this.advance()
-            break
-          }
-          this.error("expected ',' or ')'", sep.span)
-          this.syncTo(")")
-          const maybeClose = this.peek()
-          if (maybeClose.type === "punctuation" && maybeClose.value === ")") {
-            this.advance()
-          }
-          break
-        }
-        const endToken = this.previous()
-        const span = spanFrom(current.span, endToken.span)
-        const node: CallExpr = { type: "CallExpr", callee: current, args, span }
-        current = node
-        continue
-      }
-
-      break
-    }
-    return current
-  }
-
-  private parsePrimary(): Expr {
-    const token = this.peek()
-
-    if (token.type === "number") {
-      this.advance()
-      const node: Literal = {
-        type: "Literal",
-        kind: "number",
-        value: token.value,
-        span: token.span,
-      }
-      return node
-    }
-
-    if (token.type === "string") {
-      this.advance()
-      const node: Literal = {
-        type: "Literal",
-        kind: "string",
-        value: token.value,
-        span: token.span,
-      }
-      return node
-    }
-
-    if (token.type === "boolean") {
-      this.advance()
-      const node: Literal = {
-        type: "Literal",
-        kind: "boolean",
-        value: token.value,
-        span: token.span,
-      }
-      return node
-    }
-
-    if (token.type === "null") {
-      this.advance()
-      const node: Literal = { type: "Literal", kind: "null", value: null, span: token.span }
-      return node
-    }
-
-    if (token.type === "regex") {
-      this.advance()
-      const node: Literal = {
-        type: "Literal",
-        kind: "regex",
-        value: token.pattern,
-        flags: token.flags,
-        span: token.span,
-      }
-      return node
-    }
-
-    if (token.type === "identifier") {
-      this.advance()
-      const node: Identifier = { type: "Identifier", name: token.value, span: token.span }
-      return node
-    }
-
-    if (token.type === "this") {
-      this.advance()
-      const node: Identifier = { type: "Identifier", name: "this", span: token.span }
-      return node
-    }
-
-    if (token.type === "punctuation" && token.value === "(") {
-      this.advance()
-      const expr = this.parseExpression(0)
-      const closeToken = this.peek()
-      if (closeToken.type === "punctuation" && closeToken.value === ")") {
-        this.advance()
-      } else {
-        this.error("expected ')'", closeToken.span)
-        this.syncTo(")")
-        const maybeClose = this.peek()
-        if (maybeClose.type === "punctuation" && maybeClose.value === ")") {
-          this.advance()
-        }
-      }
-      return expr
-    }
-
-    if (token.type === "punctuation" && token.value === "[") {
-      return this.parseList()
-    }
-
-    this.error("unexpected token", token.span)
-    this.advance()
-    const node: ErrorExpr = { type: "ErrorExpr", message: "unexpected token", span: token.span }
-    return node
-  }
-
-  private parseList(): Expr {
-    const startToken = this.peek()
-    this.advance()
-    const elements: Expr[] = []
-    while (this.peek().type !== "eof") {
-      const next = this.peek()
-      if (next.type === "punctuation" && next.value === "]") {
-        this.advance()
-        const span = spanFrom(startToken.span, next.span)
-        const node: ListExpr = { type: "ListExpr", elements, span }
-        return node
-      }
-      const element = this.parseExpression(0)
-      elements.push(element)
-      const sep = this.peek()
-      if (sep.type === "punctuation" && sep.value === ",") {
-        this.advance()
-        const maybeClose = this.peek()
-        if (maybeClose.type === "punctuation" && maybeClose.value === "]") {
-          this.advance()
-          const span = spanFrom(startToken.span, maybeClose.span)
-          const node: ListExpr = { type: "ListExpr", elements, span }
-          return node
-        }
-        continue
-      }
-      if (sep.type === "punctuation" && sep.value === "]") {
-        this.advance()
-        const span = spanFrom(startToken.span, sep.span)
-        const node: ListExpr = { type: "ListExpr", elements, span }
-        return node
-      }
-      this.error("expected ',' or ']'", sep.span)
-      this.syncTo("]")
-      const maybeClose = this.peek()
-      if (maybeClose.type === "punctuation" && maybeClose.value === "]") {
-        const endToken = maybeClose
-        this.advance()
-        const span = spanFrom(startToken.span, endToken.span)
-        const node: ListExpr = { type: "ListExpr", elements, span }
-        return node
-      }
-      break
-    }
-    const endToken = this.previous()
-    const span = spanFrom(startToken.span, endToken.span)
-    return { type: "ListExpr", elements, span }
-  }
-
-  private error(message: string, span: Token["span"]) {
-    this.diagnostics.push({ kind: "parse", message, span })
-  }
-
-  private syncTo(value: ")" | "]") {
-    while (this.peek().type !== "eof") {
-      const token = this.peek()
-      if (token.type === "punctuation" && token.value === value) {
-        return
-      }
-      this.advance()
-    }
-  }
-
-  private peek(): Token {
-    return this.tokens[this.index]
-  }
-
-  private previous(): Token {
-    return this.tokens[Math.max(0, this.index - 1)]
-  }
-
-  private advance(): Token {
-    const token = this.tokens[this.index]
-    if (this.index < this.tokens.length - 1) this.index += 1
-    return token
-  }
-}
--- a/quartz/util/base/compiler/properties.test.ts
+++ b/quartz/util/base/compiler/properties.test.ts
@@ -1,27 +0,0 @@
-import assert from "node:assert"
-import test from "node:test"
-import { parseExpressionSource } from "./parser"
-import { buildPropertyExpressionSource } from "./properties"
-
-test("builds property expression sources", () => {
-  const cases: Array<{ input: string; expected: string }> = [
-    { input: "status", expected: "note.status" },
-    { input: "note.status", expected: "note.status" },
-    { input: "file.name", expected: "file.name" },
-    { input: "file.my-field", expected: 'file["my-field"]' },
-    { input: "my-field", expected: 'note["my-field"]' },
-    { input: 'note["my field"]', expected: 'note["my field"]' },
-    { input: "formula.total", expected: "formula.total" },
-    { input: "this.file.name", expected: "this.file.name" },
-    { input: "a.b-c.d", expected: 'note.a["b-c"].d' },
-    { input: "date(file.ctime)", expected: "date(file.ctime)" },
-  ]
-
-  for (const entry of cases) {
-    const result = buildPropertyExpressionSource(entry.input)
-    assert.strictEqual(result, entry.expected)
-    const parsed = parseExpressionSource(entry.expected)
-    assert.strictEqual(parsed.diagnostics.length, 0)
-    assert.ok(parsed.program.body)
-  }
-})
--- a/quartz/util/base/compiler/properties.ts
+++ b/quartz/util/base/compiler/properties.ts
@@ -1,27 +0,0 @@
-const simpleIdentifierPattern = /^[A-Za-z_][A-Za-z0-9_]*$/
-
-export function buildPropertyExpressionSource(property: string): string | null {
-  const trimmed = property.trim()
-  if (!trimmed) return null
-  if (trimmed.includes("(") || trimmed.includes("[") || trimmed.includes("]")) {
-    return trimmed
-  }
-  const parts = trimmed.split(".")
-  const root = parts[0]
-  const rest = parts.slice(1)
-  const buildAccess = (base: string, segments: string[]) => {
-    let source = base
-    for (const segment of segments) {
-      if (simpleIdentifierPattern.test(segment)) {
-        source = `${source}.${segment}`
-      } else {
-        source = `${source}[${JSON.stringify(segment)}]`
-      }
-    }
-    return source
-  }
-  if (root === "file" || root === "note" || root === "formula" || root === "this") {
-    return buildAccess(root, rest)
-  }
-  return buildAccess("note", parts)
-}
--- a/quartz/util/base/compiler/schema.ts
+++ b/quartz/util/base/compiler/schema.ts
@@ -1,36 +0,0 @@
-export const BUILTIN_SUMMARY_TYPES = [
-  "count",
-  "sum",
-  "average",
-  "avg",
-  "min",
-  "max",
-  "range",
-  "unique",
-  "filled",
-  "missing",
-  "median",
-  "stddev",
-  "checked",
-  "unchecked",
-  "empty",
-  "earliest",
-  "latest",
-] as const
-
-export type BuiltinSummaryType = (typeof BUILTIN_SUMMARY_TYPES)[number]
-
-export interface SummaryDefinition {
-  type: "builtin" | "formula"
-  builtinType?: BuiltinSummaryType
-  formulaRef?: string
-  expression?: string
-}
-
-export interface ViewSummaryConfig {
-  columns: Record<string, SummaryDefinition>
-}
-
-export interface PropertyConfig {
-  displayName?: string
-}
--- a/quartz/util/base/compiler/tokens.ts
+++ b/quartz/util/base/compiler/tokens.ts
@@ -1,42 +0,0 @@
-import { Span } from "./ast"
-
-export type Operator =
-  | "=="
-  | "!="
-  | ">="
-  | "<="
-  | ">"
-  | "<"
-  | "&&"
-  | "||"
-  | "+"
-  | "-"
-  | "*"
-  | "/"
-  | "%"
-  | "!"
-
-export type Punctuation = "." | "," | "(" | ")" | "[" | "]"
-
-export type NumberToken = { type: "number"; value: number; span: Span }
-export type StringToken = { type: "string"; value: string; span: Span }
-export type BooleanToken = { type: "boolean"; value: boolean; span: Span }
-export type NullToken = { type: "null"; span: Span }
-export type IdentifierToken = { type: "identifier"; value: string; span: Span }
-export type ThisToken = { type: "this"; span: Span }
-export type OperatorToken = { type: "operator"; value: Operator; span: Span }
-export type PunctuationToken = { type: "punctuation"; value: Punctuation; span: Span }
-export type RegexToken = { type: "regex"; pattern: string; flags: string; span: Span }
-export type EofToken = { type: "eof"; span: Span }
-
-export type Token =
-  | NumberToken
-  | StringToken
-  | BooleanToken
-  | NullToken
-  | IdentifierToken
-  | ThisToken
-  | OperatorToken
-  | PunctuationToken
-  | RegexToken
-  | EofToken
--- a/quartz/util/base/inspec-base.ts
+++ b/quartz/util/base/inspec-base.ts
@@ -1,278 +0,0 @@
-import yaml from "js-yaml"
-import fs from "node:fs/promises"
-import path from "node:path"
-import {
-  parseExpressionSource,
-  compileExpression,
-  buildPropertyExpressionSource,
-  BUILTIN_SUMMARY_TYPES,
-} from "./compiler"
-import { Expr, LogicalExpr, UnaryExpr, spanFrom } from "./compiler/ast"
-import { Diagnostic } from "./compiler/errors"
-
-const isRecord = (value: unknown): value is Record<string, unknown> =>
-  typeof value === "object" && value !== null && !Array.isArray(value)
-
-type CollectedExpression = {
-  kind: string
-  context: string
-  source: string
-  ast: Expr | null
-  ir: unknown
-  diagnostics: Diagnostic[]
-}
-
-const parseToExpr = (source: string, filePath: string) => {
-  const result = parseExpressionSource(source, filePath)
-  return { expr: result.program.body ?? null, diagnostics: result.diagnostics }
-}
-
-const buildLogical = (operator: "&&" | "||", expressionsList: Expr[]): Expr | null => {
-  if (expressionsList.length === 0) return null
-  let current: Expr | null = null
-  for (const next of expressionsList) {
-    if (!current) {
-      current = next
-      continue
-    }
-    const span = spanFrom(current.span, next.span)
-    const node: LogicalExpr = { type: "LogicalExpr", operator, left: current, right: next, span }
-    current = node
-  }
-  return current
-}
-
-const negateExpressions = (expressionsList: Expr[]): Expr[] =>
-  expressionsList.map((expr) => {
-    const node: UnaryExpr = {
-      type: "UnaryExpr",
-      operator: "!",
-      argument: expr,
-      span: spanFrom(expr.span, expr.span),
-    }
-    return node
-  })
-
-const buildFilterExpr = (
-  raw: unknown,
-  context: string,
-  diagnostics: Diagnostic[],
-  filePath: string,
-): Expr | null => {
-  if (typeof raw === "string") {
-    const parsed = parseToExpr(raw, filePath)
-    diagnostics.push(...parsed.diagnostics)
-    return parsed.expr
-  }
-  if (!isRecord(raw)) return null
-  if (Array.isArray(raw.and)) {
-    const parts = raw.and
-      .map((entry, index) =>
-        buildFilterExpr(entry, `${context}.and[${index}]`, diagnostics, filePath),
-      )
-      .filter((entry): entry is Expr => Boolean(entry))
-    return buildLogical("&&", parts)
-  }
-  if (Array.isArray(raw.or)) {
-    const parts = raw.or
-      .map((entry, index) =>
-        buildFilterExpr(entry, `${context}.or[${index}]`, diagnostics, filePath),
-      )
-      .filter((entry): entry is Expr => Boolean(entry))
-    return buildLogical("||", parts)
-  }
-  if (Array.isArray(raw.not)) {
-    const parts = raw.not
-      .map((entry, index) =>
-        buildFilterExpr(entry, `${context}.not[${index}]`, diagnostics, filePath),
-      )
-      .filter((entry): entry is Expr => Boolean(entry))
-    return buildLogical("&&", negateExpressions(parts))
-  }
-  return null
-}
-
-const collectPropertyExpressions = (
-  views: unknown[],
-): Map<string, { source: string; context: string }> => {
-  const entries = new Map<string, { source: string; context: string }>()
-  const addProperty = (property: string, context: string) => {
-    const key = property.trim()
-    if (!key || entries.has(key)) return
-    const source = buildPropertyExpressionSource(key)
-    if (!source) return
-    entries.set(key, { source, context })
-  }
-
-  views.forEach((view, viewIndex) => {
-    if (!isRecord(view)) return
-    const viewContext = `views[${viewIndex}]`
-    if (Array.isArray(view.order)) {
-      view.order.forEach((entry, orderIndex) => {
-        if (typeof entry === "string") {
-          addProperty(entry, `${viewContext}.order[${orderIndex}]`)
-        }
-      })
-    }
-
-    if (Array.isArray(view.sort)) {
-      view.sort.forEach((entry, sortIndex) => {
-        if (isRecord(entry) && typeof entry.property === "string") {
-          addProperty(entry.property, `${viewContext}.sort[${sortIndex}].property`)
-        }
-      })
-    }
-
-    if (typeof view.groupBy === "string") {
-      addProperty(view.groupBy, `${viewContext}.groupBy`)
-    } else if (isRecord(view.groupBy) && typeof view.groupBy.property === "string") {
-      addProperty(view.groupBy.property, `${viewContext}.groupBy.property`)
-    }
-
-    if (view.summaries && isRecord(view.summaries)) {
-      const columns =
-        "columns" in view.summaries && isRecord(view.summaries.columns)
-          ? view.summaries.columns
-          : view.summaries
-      for (const key of Object.keys(columns)) {
-        addProperty(key, `${viewContext}.summaries.${key}`)
-      }
-    }
-
-    if (typeof view.image === "string") {
-      addProperty(view.image, `${viewContext}.image`)
-    }
-
-    if (view.type === "map") {
-      const coords = typeof view.coordinates === "string" ? view.coordinates : "coordinates"
-      addProperty(coords, `${viewContext}.coordinates`)
-      if (typeof view.markerIcon === "string") {
-        addProperty(view.markerIcon, `${viewContext}.markerIcon`)
-      }
-      if (typeof view.markerColor === "string") {
-        addProperty(view.markerColor, `${viewContext}.markerColor`)
-      }
-    }
-  })
-
-  return entries
-}
-
-const main = async () => {
-  const inputPath = process.argv[2] ? String(process.argv[2]) : "content/antilibrary.base"
-  const filePath = path.resolve(process.cwd(), inputPath)
-  const raw = await fs.readFile(filePath, "utf8")
-  const parsed = yaml.load(raw)
-  const config = isRecord(parsed) ? parsed : {}
-
-  const collected: CollectedExpression[] = []
-
-  if (config.filters !== undefined) {
-    const diagnostics: Diagnostic[] = []
-    const expr = buildFilterExpr(config.filters, "filters", diagnostics, filePath)
-    collected.push({
-      kind: "filters",
-      context: "filters",
-      source: typeof config.filters === "string" ? config.filters : JSON.stringify(config.filters),
-      ast: expr,
-      ir: expr ? compileExpression(expr) : null,
-      diagnostics,
-    })
-  }
-
-  if (isRecord(config.formulas)) {
-    for (const [name, value] of Object.entries(config.formulas)) {
-      if (typeof value !== "string") continue
-      const parsedExpr = parseToExpr(value, filePath)
-      collected.push({
-        kind: "formula",
-        context: `formulas.${name}`,
-        source: value,
-        ast: parsedExpr.expr,
-        ir: parsedExpr.expr ? compileExpression(parsedExpr.expr) : null,
-        diagnostics: parsedExpr.diagnostics,
-      })
-    }
-  }
-
-  const topLevelSummaries = isRecord(config.summaries) ? config.summaries : {}
-
-  if (isRecord(config.summaries)) {
-    for (const [name, value] of Object.entries(config.summaries)) {
-      if (typeof value !== "string") continue
-      const parsedExpr = parseToExpr(value, filePath)
-      collected.push({
-        kind: "summary",
-        context: `summaries.${name}`,
-        source: value,
-        ast: parsedExpr.expr,
-        ir: parsedExpr.expr ? compileExpression(parsedExpr.expr) : null,
-        diagnostics: parsedExpr.diagnostics,
-      })
-    }
-  }
-
-  if (Array.isArray(config.views)) {
-    config.views.forEach((view, index) => {
-      if (!isRecord(view)) return
-      if (view.filters !== undefined) {
-        const diagnostics: Diagnostic[] = []
-        const expr = buildFilterExpr(view.filters, `views[${index}].filters`, diagnostics, filePath)
-        collected.push({
-          kind: "view.filter",
-          context: `views[${index}].filters`,
-          source: typeof view.filters === "string" ? view.filters : JSON.stringify(view.filters),
-          ast: expr,
-          ir: expr ? compileExpression(expr) : null,
-          diagnostics,
-        })
-      }
-
-      if (view.summaries && isRecord(view.summaries)) {
-        const columns =
-          "columns" in view.summaries && isRecord(view.summaries.columns)
-            ? view.summaries.columns
-            : view.summaries
-        for (const [column, summaryValue] of Object.entries(columns)) {
-          if (typeof summaryValue !== "string") continue
-          const normalized = summaryValue.toLowerCase().trim()
-          const builtins = new Set<string>(BUILTIN_SUMMARY_TYPES)
-          if (builtins.has(normalized)) continue
-          const summarySource =
-            summaryValue in topLevelSummaries && typeof topLevelSummaries[summaryValue] === "string"
-              ? String(topLevelSummaries[summaryValue])
-              : summaryValue
-          const parsedExpr = parseToExpr(summarySource, filePath)
-          collected.push({
-            kind: "view.summary",
-            context: `views[${index}].summaries.${column}`,
-            source: summarySource,
-            ast: parsedExpr.expr,
-            ir: parsedExpr.expr ? compileExpression(parsedExpr.expr) : null,
-            diagnostics: parsedExpr.diagnostics,
-          })
-        }
-      }
-    })
-  }
-
-  const views = Array.isArray(config.views) ? config.views : []
-  const propertyExpressions = collectPropertyExpressions(views)
-  for (const [_, entry] of propertyExpressions.entries()) {
-    const parsedExpr = parseToExpr(entry.source, filePath)
-    collected.push({
-      kind: "property",
-      context: entry.context,
-      source: entry.source,
-      ast: parsedExpr.expr,
-      ir: parsedExpr.expr ? compileExpression(parsedExpr.expr) : null,
-      diagnostics: parsedExpr.diagnostics,
-    })
-  }
-
-  const payload = { file: inputPath, count: collected.length, expressions: collected }
-
-  process.stdout.write(JSON.stringify(payload, null, 2))
-}
-
-main()
--- a/quartz/util/base/query.ts
+++ b/quartz/util/base/query.ts
@@ -1,248 +0,0 @@
-import { QuartzPluginData } from "../../plugins/vfile"
-import { evaluateSummaryExpression, valueToUnknown, EvalContext, ProgramIR } from "./compiler"
-import { SummaryDefinition, ViewSummaryConfig, BuiltinSummaryType } from "./types"
-
-type SummaryValueResolver = (
-  file: QuartzPluginData,
-  column: string,
-  allFiles: QuartzPluginData[],
-) => unknown
-
-type SummaryContextFactory = (file: QuartzPluginData) => EvalContext
-
-export function computeColumnSummary(
-  column: string,
-  files: QuartzPluginData[],
-  summary: SummaryDefinition,
-  allFiles: QuartzPluginData[] = [],
-  valueResolver: SummaryValueResolver,
-  getContext: SummaryContextFactory,
-  summaryExpression?: ProgramIR,
-): string | number | undefined {
-  if (files.length === 0) {
-    return undefined
-  }
-
-  const values = files.map((file) => valueResolver(file, column, allFiles))
-
-  if (summary.type === "builtin" && summary.builtinType) {
-    return computeBuiltinSummary(values, summary.builtinType)
-  }
-
-  if (summary.type === "formula" && summary.expression) {
-    if (summaryExpression) {
-      const summaryCtx = getContext(files[0])
-      summaryCtx.diagnosticContext = `summaries.${column}`
-      summaryCtx.diagnosticSource = summary.expression
-      summaryCtx.rows = files
-      const value = evaluateSummaryExpression(summaryExpression, values, summaryCtx)
-      const unknownValue = valueToUnknown(value)
-      if (typeof unknownValue === "number" || typeof unknownValue === "string") {
-        return unknownValue
-      }
-      return undefined
-    }
-  }
-
-  return undefined
-}
-
-function computeBuiltinSummary(
-  values: any[],
-  type: BuiltinSummaryType,
-): string | number | undefined {
-  switch (type) {
-    case "count":
-      return values.length
-
-    case "sum": {
-      const nums = values.filter((v) => typeof v === "number")
-      if (nums.length === 0) return undefined
-      return nums.reduce((acc, v) => acc + v, 0)
-    }
-
-    case "average":
-    case "avg": {
-      const nums = values.filter((v) => typeof v === "number")
-      if (nums.length === 0) return undefined
-      const sum = nums.reduce((acc, v) => acc + v, 0)
-      return Math.round((sum / nums.length) * 100) / 100
-    }
-
-    case "min": {
-      const comparable = values.filter(
-        (v) => typeof v === "number" || v instanceof Date || typeof v === "string",
-      )
-      if (comparable.length === 0) return undefined
-      const normalized = comparable.map((v) => (v instanceof Date ? v.getTime() : v))
-      const min = Math.min(...normalized.filter((v) => typeof v === "number"))
-      if (isNaN(min)) {
-        const strings = comparable.filter((v) => typeof v === "string") as string[]
-        if (strings.length === 0) return undefined
-        return strings.sort()[0]
-      }
-      if (comparable.some((v) => v instanceof Date)) {
-        return new Date(min).toISOString().split("T")[0]
-      }
-      return min
-    }
-
-    case "max": {
-      const comparable = values.filter(
-        (v) => typeof v === "number" || v instanceof Date || typeof v === "string",
-      )
-      if (comparable.length === 0) return undefined
-      const normalized = comparable.map((v) => (v instanceof Date ? v.getTime() : v))
-      const max = Math.max(...normalized.filter((v) => typeof v === "number"))
-      if (isNaN(max)) {
-        const strings = comparable.filter((v) => typeof v === "string") as string[]
-        if (strings.length === 0) return undefined
-        return strings.sort().reverse()[0]
-      }
-      if (comparable.some((v) => v instanceof Date)) {
-        return new Date(max).toISOString().split("T")[0]
-      }
-      return max
-    }
-
-    case "range": {
-      const comparable = values.filter(
-        (v) => typeof v === "number" || v instanceof Date || typeof v === "string",
-      )
-      if (comparable.length === 0) return undefined
-      const normalized = comparable.map((v) => (v instanceof Date ? v.getTime() : v))
-      const nums = normalized.filter((v) => typeof v === "number")
-      if (nums.length === 0) return undefined
-      const min = Math.min(...nums)
-      const max = Math.max(...nums)
-      if (comparable.some((v) => v instanceof Date)) {
-        return `${new Date(min).toISOString().split("T")[0]} - ${new Date(max).toISOString().split("T")[0]}`
-      }
-      return `${min} - ${max}`
-    }
-
-    case "unique": {
-      const nonNull = values.filter((v) => v !== undefined && v !== null && v !== "")
-      const unique = new Set(nonNull.map((v) => (v instanceof Date ? v.toISOString() : String(v))))
-      return unique.size
-    }
-
-    case "filled": {
-      const filled = values.filter((v) => v !== undefined && v !== null && v !== "")
-      return filled.length
-    }
-
-    case "missing": {
-      const missing = values.filter((v) => v === undefined || v === null || v === "")
-      return missing.length
-    }
-
-    case "median": {
-      const nums = values.filter((v) => typeof v === "number") as number[]
-      if (nums.length === 0) return undefined
-      const sorted = [...nums].sort((a, b) => a - b)
-      const mid = Math.floor(sorted.length / 2)
-      if (sorted.length % 2 === 0) {
-        return (sorted[mid - 1] + sorted[mid]) / 2
-      }
-      return sorted[mid]
-    }
-
-    case "stddev": {
-      const nums = values.filter((v) => typeof v === "number") as number[]
-      if (nums.length === 0) return undefined
-      const mean = nums.reduce((acc, v) => acc + v, 0) / nums.length
-      const variance = nums.reduce((acc, v) => acc + (v - mean) * (v - mean), 0) / nums.length
-      return Math.round(Math.sqrt(variance) * 100) / 100
-    }
-
-    case "checked":
-      return values.filter((v) => v === true).length
-
-    case "unchecked":
-      return values.filter((v) => v === false).length
-
-    case "empty": {
-      const count = values.filter(
-        (v) =>
-          v === undefined ||
-          v === null ||
-          v === "" ||
-          (Array.isArray(v) && v.length === 0) ||
-          (typeof v === "object" && v !== null && !Array.isArray(v) && Object.keys(v).length === 0),
-      ).length
-      return count
-    }
-
-    case "earliest": {
-      const dates = values.filter(
-        (v) =>
-          v instanceof Date ||
-          (typeof v === "string" && /^\d{4}-\d{2}-\d{2}/.test(v)) ||
-          typeof v === "number",
-      )
-      if (dates.length === 0) return undefined
-      const timestamps = dates.map((v) => {
-        if (v instanceof Date) return v.getTime()
-        if (typeof v === "string") return new Date(v).getTime()
-        return v
-      })
-      const earliest = Math.min(...timestamps)
-      return new Date(earliest).toISOString().split("T")[0]
-    }
-
-    case "latest": {
-      const dates = values.filter(
-        (v) =>
-          v instanceof Date ||
-          (typeof v === "string" && /^\d{4}-\d{2}-\d{2}/.test(v)) ||
-          typeof v === "number",
-      )
-      if (dates.length === 0) return undefined
-      const timestamps = dates.map((v) => {
-        if (v instanceof Date) return v.getTime()
-        if (typeof v === "string") return new Date(v).getTime()
-        return v
-      })
-      const latest = Math.max(...timestamps)
-      return new Date(latest).toISOString().split("T")[0]
-    }
-
-    default:
-      return undefined
-  }
-}
-
-export function computeViewSummaries(
-  columns: string[],
-  files: QuartzPluginData[],
-  summaryConfig: ViewSummaryConfig | undefined,
-  allFiles: QuartzPluginData[] = [],
-  getContext: SummaryContextFactory,
-  valueResolver: SummaryValueResolver,
-  summaryExpressions?: Record<string, ProgramIR>,
-): Record<string, string | number | undefined> {
-  const results: Record<string, string | number | undefined> = {}
-
-  if (!summaryConfig?.columns) {
-    return results
-  }
-
-  for (const column of columns) {
-    const summary = summaryConfig.columns[column]
-    if (summary) {
-      const expression = summaryExpressions ? summaryExpressions[column] : undefined
-      results[column] = computeColumnSummary(
-        column,
-        files,
-        summary,
-        allFiles,
-        valueResolver,
-        getContext,
-        expression,
-      )
-    }
-  }
-
-  return results
-}
--- a/quartz/util/base/render.ts
+++ b/quartz/util/base/render.ts
--- a/quartz/util/base/types.test.ts
+++ b/quartz/util/base/types.test.ts
@@ -1,31 +0,0 @@
-import assert from "node:assert"
-import test from "node:test"
-import { parseViews, parseViewSummaries } from "./types"
-
-test("parseViews preserves raw filters", () => {
-  const views = parseViews([
-    { type: "table", name: "test", filters: 'status == "done"', order: ["file.name"] },
-  ])
-
-  assert.strictEqual(views.length, 1)
-  assert.strictEqual(views[0].filters, 'status == "done"')
-  assert.deepStrictEqual(views[0].order, ["file.name"])
-})
-
-test("parseViews rejects missing type/name", () => {
-  assert.throws(() => parseViews([{}]))
-})
-
-test("parseViewSummaries resolves builtin and formula refs", () => {
-  const summaries = parseViewSummaries(
-    { price: "Average", score: "avgScore", extra: "values.length" },
-    { avgScore: "values.mean()" },
-  )
-
-  assert.ok(summaries)
-  if (!summaries) return
-  assert.strictEqual(summaries.columns.price.type, "builtin")
-  assert.strictEqual(summaries.columns.score.type, "formula")
-  assert.strictEqual(summaries.columns.score.formulaRef, "avgScore")
-  assert.strictEqual(summaries.columns.extra.type, "formula")
-})
--- a/quartz/util/base/types.ts
+++ b/quartz/util/base/types.ts
@@ -1,119 +0,0 @@
-import {
-  SummaryDefinition,
-  ViewSummaryConfig,
-  PropertyConfig,
-  BuiltinSummaryType,
-  BUILTIN_SUMMARY_TYPES,
-} from "./compiler/schema"
-
-export type { SummaryDefinition, ViewSummaryConfig, PropertyConfig, BuiltinSummaryType }
-export { BUILTIN_SUMMARY_TYPES }
-
-const isRecord = (value: unknown): value is Record<string, unknown> =>
-  typeof value === "object" && value !== null && !Array.isArray(value)
-
-const isNonEmptyString = (value: unknown): value is string =>
-  typeof value === "string" && value.trim().length > 0
-
-export type BaseFileFilter =
-  | string
-  | { and: BaseFileFilter[] }
-  | { or: BaseFileFilter[] }
-  | { not: BaseFileFilter[] }
-
-export interface BaseFile {
-  filters?: BaseFileFilter
-  views: BaseView[]
-  properties?: Record<string, PropertyConfig>
-  summaries?: Record<string, string>
-  formulas?: Record<string, string>
-}
-
-export interface BaseView {
-  type: "table" | "list" | "gallery" | "board" | "calendar" | "card" | "cards" | "map"
-  name: string
-  order?: string[]
-  sort?: BaseSortConfig[]
-  columnSize?: Record<string, number>
-  groupBy?: string | BaseGroupBy
-  limit?: number
-  filters?: BaseFileFilter
-  summaries?: Record<string, string> | ViewSummaryConfig
-  image?: string
-  cardSize?: number
-  cardAspect?: number
-  nestedProperties?: boolean
-  indentProperties?: boolean
-  separator?: string
-  date?: string
-  dateField?: string
-  dateProperty?: string
-  coordinates?: string
-  markerIcon?: string
-  markerColor?: string
-  defaultZoom?: number
-  defaultCenter?: [number, number]
-  clustering?: boolean
-  groupSizes?: Record<string, number>
-  groupAspects?: Record<string, number>
-}
-
-export interface BaseSortConfig {
-  property: string
-  direction: "ASC" | "DESC"
-}
-
-export interface BaseGroupBy {
-  property: string
-  direction: "ASC" | "DESC"
-}
-
-export function parseViews(raw: unknown[]): BaseView[] {
-  return raw.map((entry) => {
-    if (!isRecord(entry)) throw new Error("Each view must be an object")
-    const { type, name } = entry
-    if (!isNonEmptyString(type) || !isNonEmptyString(name)) {
-      throw new Error("Each view must have 'type' and 'name' fields")
-    }
-    return { ...entry, type, name } as BaseView
-  })
-}
-
-export function parseViewSummaries(
-  viewSummaries: Record<string, string> | ViewSummaryConfig | undefined,
-  topLevelSummaries?: Record<string, string>,
-): ViewSummaryConfig | undefined {
-  if (!viewSummaries || typeof viewSummaries !== "object") return undefined
-
-  if ("columns" in viewSummaries && typeof viewSummaries.columns === "object") {
-    return viewSummaries as ViewSummaryConfig
-  }
-
-  const columns: Record<string, SummaryDefinition> = {}
-
-  for (const [column, summaryValue] of Object.entries(viewSummaries)) {
-    if (typeof summaryValue !== "string") continue
-
-    const normalized = summaryValue.toLowerCase().trim()
-
-    if (BUILTIN_SUMMARY_TYPES.includes(normalized as BuiltinSummaryType)) {
-      columns[column] = { type: "builtin", builtinType: normalized as BuiltinSummaryType }
-      continue
-    }
-
-    if (topLevelSummaries && summaryValue in topLevelSummaries) {
-      columns[column] = {
-        type: "formula",
-        formulaRef: summaryValue,
-        expression: topLevelSummaries[summaryValue],
-      }
-      continue
-    }
-
-    if (summaryValue.includes("(") || summaryValue.includes(".")) {
-      columns[column] = { type: "formula", expression: summaryValue }
-    }
-  }
-
-  return Object.keys(columns).length > 0 ? { columns } : undefined
-}
--- a/quartz/util/glob.ts
+++ b/quartz/util/glob.ts
@@ -10,12 +10,13 @@ export async function glob(
  pattern: string,
  cwd: string,
  ignorePatterns: string[],
+  respectGitignore: boolean = true,
 ): Promise<FilePath[]> {
  const fps = (
    await globby(pattern, {
      cwd,
      ignore: ignorePatterns,
-      gitignore: true,
+      gitignore: respectGitignore,
    })
  ).map(toPosixPath)
  return fps as FilePath[]
--- a/quartz/util/path.ts
+++ b/quartz/util/path.ts
@@ -73,7 +73,7 @@ export function slugifyFilePath(fp: FilePath, excludeExt?: boolean): FullSlug {
  fp = stripSlashes(fp) as FilePath
  let ext = getFileExtension(fp)
  const withoutFileExt = fp.replace(new RegExp(ext + "$"), "")
-  if (excludeExt || [".md", ".html", ".base", undefined].includes(ext)) {
+  if (excludeExt || [".md", ".html", undefined].includes(ext)) {
    ext = ""
  }

--- a/quartz/util/wikilinks.ts
+++ b/quartz/util/wikilinks.ts
@@ -1,94 +0,0 @@
-import { FilePath, FullSlug, slugifyFilePath } from "./path"
-
-export type WikilinkWithPosition = {
-  wikilink: ParsedWikilink
-  start: number
-  end: number
-}
-
-export type ParsedWikilink = {
-  raw: string
-  target: string
-  anchor?: string
-  alias?: string
-  embed: boolean
-}
-
-export type ResolvedWikilink = {
-  slug: FullSlug
-  anchor?: string
-}
-
-const wikilinkRegex = /^!?\[\[([^\]|#]+)(?:#([^\]|]+))?(?:\|([^\]]+))?\]\]$/
-
-export function parseWikilink(text: string): ParsedWikilink | null {
-  const trimmed = text.trim()
-  const match = wikilinkRegex.exec(trimmed)
-  if (!match) return null
-
-  const [, target, anchor, alias] = match
-  return {
-    raw: trimmed,
-    target: target?.trim() ?? "",
-    anchor: anchor?.trim(),
-    alias: alias?.trim(),
-    embed: trimmed.startsWith("!"),
-  }
-}
-
-export function resolveWikilinkTarget(
-  parsed: ParsedWikilink,
-  currentSlug: FullSlug,
-): ResolvedWikilink | null {
-  const target = parsed.target.trim()
-  if (!target) return null
-
-  if (target.startsWith("/")) {
-    const slug = slugifyFilePath(target.slice(1).replace(/\\/g, "/") as FilePath)
-    return { slug, anchor: parsed.anchor }
-  }
-
-  const currentParts = currentSlug.split("/")
-  const currentDir = currentParts.slice(0, -1)
-
-  const targetParts = target.replace(/\\/g, "/").split("/")
-  const resolved: string[] = [...currentDir]
-
-  for (const part of targetParts) {
-    if (part === "..") {
-      resolved.pop()
-    } else if (part !== "." && part.length > 0) {
-      resolved.push(part)
-    }
-  }
-
-  const slug = slugifyFilePath(resolved.join("/") as FilePath)
-  return { slug, anchor: parsed.anchor }
-}
-
-const globalWikilinkRegex = /!?\[\[([^\]|#]+)(?:#([^\]|]+))?(?:\|([^\]]+))?\]\]/g
-
-export function extractWikilinksWithPositions(text: string): WikilinkWithPosition[] {
-  const results: WikilinkWithPosition[] = []
-  let match: RegExpExecArray | null
-
-  globalWikilinkRegex.lastIndex = 0
-
-  while ((match = globalWikilinkRegex.exec(text)) !== null) {
-    const [fullMatch, target, anchor, alias] = match
-
-    results.push({
-      wikilink: {
-        raw: fullMatch,
-        target: target?.trim() ?? "",
-        anchor: anchor?.trim(),
-        alias: alias?.trim(),
-        embed: fullMatch.startsWith("!"),
-      },
-      start: match.index,
-      end: match.index + fullMatch.length,
-    })
-  }
-
-  return results
-}
Author	SHA1	Message	Date
Ignacio Ballesteros	678fb315d3	feat: unified org-garden	2026-02-21 13:10:30 +01:00
Ignacio Ballesteros	a4582230b5	feat: unified watch server under org-garden	2026-02-21 00:36:31 +01:00
Ignacio Ballesteros	1076bf31ed	feat: pipeline - org2md file watch	2026-02-20 23:31:06 +01:00
Ignacio Ballesteros	dc348185a7	feat(pipeline): refactor into its own project	2026-02-20 17:54:12 +01:00
Ignacio Ballesteros	0ea5808cd2	initial flake app	2026-02-20 15:08:04 +01:00
Ignacio Ballesteros	d913138726	fix(ox-hugo): serve gitignored content and static assets, fix figure shortcode order - glob.ts: add optional respectGitignore param (default true) so callers can opt out of gitignore filtering for generated directories - build.ts: pass respectGitignore=false when globbing content/ so gitignored-but-generated markdown files are always picked up - static.ts: copy top-level static/ → output/ root with gitignore disabled, mirroring Hugo's convention so ox-hugo images at static/ox-hugo/* are served at /ox-hugo/* as expected by the exported markdown src paths - oxhugofm.ts: run replaceFigureWithMdImg before removeHugoShortcode so the precise figure regex sees the intact shortcode before the generic {{.*}} stripper destroys it; also upgrade figureTagRegex to figureShortcodeRegex which matches the full shortcode form ox-hugo actually emits	2026-02-20 11:11:42 +01:00
Ignacio Ballesteros	3deec2d011	example: images	2026-02-20 10:51:05 +01:00
Ignacio Ballesteros	a907d6513b	docs(examples): add image reference example note with all three scenarios Adds notes/example-images.org demonstrating how org-mode image links of each kind pass through ox-hugo and land in content/: Scenario 1 — External URL: linked image using #+ATTR_HTML + a local thumbnail as the display image; ox-hugo emits a {{< figure >}} shortcode with a link= attribute pointing to the remote URL. Scenario 2 — Local file (same notes dir): [[file:quartz-logo.jpg]] ox-hugo copies the file to static/ox-hugo/ and rewrites the src to /ox-hugo/quartz-logo.jpg. Scenario 3 — Absolute path outside notes dir: [[file:/abs/path/img.jpg]] ox-hugo copies the file from outside the notes tree into static/ox-hugo/ and rewrites the src identically to scenario 2. Also adds: notes/quartz-logo.jpg — image asset for scenarios 1 & 2 notes-external/ — simulates an assets dir outside notes static/ox-hugo/ added to .gitignore (populated at export time)	2026-02-20 10:09:54 +01:00
Ignacio Ballesteros	9ecc9336da	docs: update AGENTS.md and README with pipeline and citation details	2026-02-20 10:02:36 +01:00
Ignacio Ballesteros	511b003da8	Add Elixir markdown pipeline with org-citar citation resolution Introduces scripts/pipeline/, a Mix project that runs as a post-export transformation pass over content/*.md before Quartz builds the site. Pipeline (scripts/export.exs phase 3): - Compiles and loads the Mix project at export time (cached after first run) - Applies a list of Transform modules sequentially over all .md files - Only rewrites files that were actually changed Citations transform (Pipeline.Transforms.Citations): - Resolves [cite:@key] and bare cite:key syntax produced by ox-hugo/citar - Resolution chain: Zotero BBT JSON-RPC → BibTeX file → DOI/bare-key fallback - Zotero probe uses a no-op JSON-RPC call (cayw endpoint blocks indefinitely) - Zotero resolver fetches PDF attachments via item.attachments, producing zotero://open-pdf/... links; falls back to zotero://select/library/items/... - BibTeX resolver parses .bib files with a simple regex parser (no deps) - DOI resolver is the always-succeeding last resort Configuration via env vars: BIBTEX_FILE — path to .bib file for fallback resolution ZOTERO_URL — Zotero base URL (default: http://localhost:23119) CITATION_MODE — silent \| warn (default) \| strict Adding future transforms requires only implementing Pipeline.Transform behaviour and appending the module to the transforms list in export.exs.	2026-02-20 10:00:11 +01:00
Ignacio Ballesteros	ec1426241c	Ignore generated content/ output and erl_crash.dump	2026-02-19 18:21:02 +01:00
Ignacio Ballesteros	692f23bc36	Add org-roam workflow: ox-hugo export pipeline and example notes - Add flake.nix dev shell with Node.js 22, Elixir, and Emacs+ox-hugo - Add scripts/export.exs: exports org-roam notes to content/ via ox-hugo, mirroring subdirectory structure and generating a fallback index.md - Add npm scripts: export, build:notes, serve:notes - Configure FrontMatter plugin for TOML (ox-hugo default output) - Replace ObsidianFlavoredMarkdown with OxHugoFlavouredMarkdown - Add example notes: Madrid public transport (metro, bus, roads) - Update README and AGENTS.md with org-roam workflow instructions	2026-02-19 18:20:43 +01:00
Ignacio Ballesteros	86eaaae945	Add branch workflow documentation to AGENTS.md	2026-02-18 12:44:49 +01:00