wiki/README.md

markdown "sync" framework

An open-source publishing framework built for AI agents and developers to ship websites, docs, or blogs. Write markdown, sync from the terminal. Your content is instantly available to browsers, LLMs, and AI agents. Built on Convex and Netlify.

Write markdown locally, run npm run sync (dev) or npm run sync:prod (production), and content appears instantly across all connected browsers. Built with React, Convex, and Vite. Optimized for AEO, GEO, and LLM discovery.

How publishing works: Write posts in markdown, run npm run sync for development or npm run sync:prod for production, and they appear on your live site immediately. No rebuild or redeploy needed. Convex handles real-time data sync, so all connected browsers update automatically.

Sync commands:

Development:

• npm run sync - Sync markdown content
• npm run sync:discovery - Update discovery files (AGENTS.md, llms.txt)
• npm run sync:all - Sync content + discovery files together

Production:

• npm run sync:prod - Sync markdown content
• npm run sync:discovery:prod - Update discovery files
• npm run sync:all:prod - Sync content + discovery files together

How versioning works: Markdown files live in content/blog/ and content/pages/. These are regular files in your git repo. Commit changes, review diffs, roll back like any codebase. The sync command pushes content to Convex.

# Edit, commit, sync
git add content/blog/my-post.md
git commit -m "Update post"
npm run sync        # dev
npm run sync:prod   # production

MCP Server: The site includes an HTTP-based Model Context Protocol (MCP) server at /mcp for AI tool integration. Connect Cursor, Claude Desktop, and other MCP clients to access blog content programmatically. See How to Use the MCP Server for setup instructions.

Documentation

• Setup Guide - Complete fork and deployment guide
• Configuration Guide - Automated or manual fork setup
• Full Documentation - Docs for all features and configuration

AI Development Tools

The project includes documentation optimized for AI coding assistants:

• CLAUDE.md - Project instructions for Claude Code CLI with workflows, commands, and conventions
• AGENTS.md - General AI agent instructions for understanding the codebase structure
• .claude/skills/ - Focused skill documentation:
  • frontmatter.md - Complete frontmatter syntax and all field options
  • convex.md - Convex patterns specific to this app
  • sync.md - How sync commands work and content flow

These files are automatically updated during npm run sync:discovery with current site statistics.

Fork Configuration

After forking this project, you have two options to configure your site:

Option 1: Automated (Recommended)

Run a single command to configure all files automatically:

# Copy the example config
cp fork-config.json.example fork-config.json

# Edit with your site information
# Open fork-config.json and update the values

# Apply all changes
npm run configure

This updates all 11 configuration files in one step.

Option 2: Manual

Follow the step-by-step guide in FORK_CONFIG.md to update each file manually. This guide includes code snippets and an AI agent prompt for assistance.

Features

• Markdown-based blog posts with frontmatter
• Syntax highlighting for code blocks
• Four theme options: Dark, Light, Tan (default), Cloud
• Real-time data with Convex
• Fully responsive design
• Real-time analytics at /stats with visitor map
• Full text search with Command+K shortcut
• Featured section with list/card view toggle
• Logo gallery with continuous marquee scroll or static grid
• GitHub contributions graph with year navigation
• Static raw markdown files at /raw/{slug}.md
• Dedicated blog page with configurable navigation order and featured layout
• Markdown writing page at /write with frontmatter reference
• AI Agent chat (powered by Anthropic Claude) on Write page and optionally in right sidebar
• Tag pages at /tags/[tag] with view mode toggle
• Related posts based on shared tags
• Footer component with markdown support and images
• Social footer with customizable social links and copyright
• Right sidebar for CopyPageDropdown and AI chat
• Contact forms on any page or post
• Newsletter subscriptions and admin UI
• Homepage post limit with optional "read more" link
• Blog page featured layout with hero post
• Show image at top of posts/pages
• Version control with 3-day history for posts and pages (diff view, one-click restore)

SEO and Discovery

• RSS feeds at /rss.xml and /rss-full.xml (with full content)
• Dynamic sitemap at /sitemap.xml
• JSON-LD structured data for Google rich results
• Open Graph and Twitter Card meta tags
• robots.txt with AI crawler rules
• llms.txt for AI agent discovery

AI and LLM Access

• /api/posts - JSON list of all posts for agents
• /api/post?slug=xxx - Single post JSON or markdown
• /api/export - Batch export all posts with full content
• /raw/{slug}.md - Static raw markdown files for each post and page
• /rss-full.xml - Full content RSS for LLM ingestion
• /.well-known/ai-plugin.json - AI plugin manifest
• /openapi.yaml - OpenAPI 3.0 specification
• Copy Page dropdown for sharing to ChatGPT, Claude, Perplexity (uses raw markdown URLs for better AI parsing)

MCP Server

The site includes an HTTP-based Model Context Protocol (MCP) server for AI tool integration. It allows AI assistants like Cursor and Claude Desktop to access blog content programmatically.

Endpoint: https://www.markdown.fast/mcp

Features:

• 24/7 availability via Netlify Edge Functions (no local machine required)
• Public access with rate limiting (50 req/min per IP)
• Optional API key for higher limits (1000 req/min)
• Seven tools: list_posts, get_post, list_pages, get_page, get_homepage, search_content, export_all

Configuration:

Add to your Cursor config (~/.cursor/mcp.json):

{
  "mcpServers": {
    "markdown-fast": {
      "url": "https://www.markdown.fast/mcp"
    }
  }
}

For forks: Set VITE_CONVEX_URL in Netlify environment variables. Optionally set MCP_API_KEY for authenticated access.

See How to Use the MCP Server for full documentation.

Content Import

• Import external URLs as markdown posts using Firecrawl
• Run npm run import <url> to scrape and create draft posts locally
• Then sync to dev or prod with npm run sync or npm run sync:prod

Dashboard

The framework includes a centralized dashboard at /dashboard for managing content and configuring your site. Features include:

• Content management: Edit posts and pages with live preview
• Version control: View history, compare diffs, restore previous versions (3-day retention)
• Sync commands: Run sync operations from the browser
• Site configuration: Configure all settings via UI
• Newsletter management: Integrated subscriber and email management
• AI Agent: Writing assistance powered by Claude
• Analytics: Real-time stats dashboard

WorkOS authentication is recommended so no one has access to your dashboard if it's enabled. Configure it in siteConfig.ts to protect the dashboard in production. See How to use the Markdown sync dashboard and How to setup WorkOS for details.

Newsletter and Email

The framework includes AgentMail integration for newsletter subscriptions and contact forms. Features include:

• Newsletter subscriptions and sending
• Contact forms on any post or page
• Automated weekly digests (Sundays 9am UTC)
• Developer notifications (new subscriber alerts, weekly stats summaries)
• Admin UI for subscriber management at /newsletter-admin
• CLI tools for sending newsletters and stats
• Custom email composition with markdown support
• Email statistics dashboard

See the AgentMail setup guide for configuration instructions.

Getting Started

Prerequisites

• Node.js 18 or higher
• A Convex account

Setup

1. Install dependencies:

npm install

2. Initialize Convex:

npx convex dev

This will create your Convex project and generate the .env.local file.

3. Start the development server:

npm run dev

4. Open http://localhost:5173

Writing Blog Posts

Create markdown files in content/blog/ with frontmatter:

Static Pages (Optional)

Create optional pages like About, Projects, or Contact in content/pages/:

---
title: "About"
slug: "about"
published: true
order: 1
---

Your page content here...

Pages appear as navigation links in the top right, next to the theme toggle. The order field controls display order (lower numbers first).

---
title: "Your Post Title"
description: "A brief description"
date: "2025-01-15"
slug: "your-post-slug"
published: true
tags: ["tag1", "tag2"]
readTime: "5 min read"
image: "/images/my-header.png"
excerpt: "Short text for featured cards"
---

Your markdown content here...

Logo Gallery

The homepage includes a scrolling logo gallery with sample logos. Configure in siteConfig:

Disable the gallery

logoGallery: {
  enabled: false,
  // ...
},

GitHub Contributions Graph

Display your GitHub contribution activity on the homepage. Configure in src/config/siteConfig.ts:

gitHubContributions: {
  enabled: true,           // Set to false to hide
  username: "yourusername", // Your GitHub username
  showYearNavigation: true, // Show arrows to navigate between years
  linkToProfile: true,      // Click graph to open GitHub profile
  title: "GitHub Activity", // Optional title above the graph
},

Visitor Map

Display real-time visitor locations on a world map on the stats page. Uses Netlify's built-in geo detection (no third-party API needed). Privacy friendly: only stores city, country, and coordinates. No IP addresses stored.

Configure in src/config/siteConfig.ts:

visitorMap: {
  enabled: true,        // Set to false to hide the visitor map
  title: "Live Visitors", // Optional title above the map
},

Syncing Posts

Posts are synced to Convex. The sync script reads markdown files from content/blog/ and content/pages/, then uploads them to your Convex database.

Environment Files

File	Purpose
.env.local	Development deployment URL (created by npx convex dev)
.env.production.local	Production deployment URL (create manually)

Both files are gitignored. Each developer creates their own.

Sync Commands

Development:

• npm run sync - Sync markdown content to development Convex
• npm run sync:discovery - Update discovery files (AGENTS.md, llms.txt) with development data
• npm run sync:all - Sync content + discovery files together

Production:

• npm run sync:prod - Sync markdown content to production Convex
• npm run sync:discovery:prod - Update discovery files with production data
• npm run sync:all:prod - Sync content + discovery files together

Development sync:

npm run sync

Production sync:

First, create .env.production.local with your production Convex URL:

VITE_CONVEX_URL=https://your-prod-deployment.convex.cloud

Then sync:

npm run sync:prod

Deployment

Netlify

For detailed setup, see the Convex Netlify Deployment Guide.

1. Deploy Convex functions to production:

npx convex deploy

Note the production URL (e.g., https://your-deployment.convex.cloud).

2. Connect your repository to Netlify
3. Configure build settings:
   • Build command: npm ci --include=dev && npx convex deploy --cmd 'npm run build'
   • Publish directory: dist
4. Add environment variables in Netlify dashboard:
   • CONVEX_DEPLOY_KEY - Generate from Convex Dashboard > Project Settings > Deploy Key
   • VITE_CONVEX_URL - Your production Convex URL (e.g., https://your-deployment.convex.cloud)

The CONVEX_DEPLOY_KEY deploys functions at build time. The VITE_CONVEX_URL is required for edge functions (RSS, sitemap, API) to proxy requests at runtime.

Build issues? Netlify sets NODE_ENV=production which skips devDependencies. The --include=dev flag fixes this. See netlify-deploy-fix.md for detailed troubleshooting.

Project Structure

markdown-site/
├── content/blog/      # Markdown blog posts
├── convex/            # Convex backend
│   ├── http.ts        # HTTP endpoints (sitemap, API, RSS)
│   ├── posts.ts       # Post queries and mutations
│   ├── rss.ts         # RSS feed generation
│   └── schema.ts      # Database schema
├── netlify/           # Netlify edge functions
│   └── edge-functions/
│       ├── rss.ts     # RSS feed proxy
│       ├── sitemap.ts # Sitemap proxy
│       ├── api.ts     # API endpoint proxy
│       └── botMeta.ts # OG crawler detection
├── public/            # Static assets
│   ├── images/        # Blog images and OG images
│   ├── robots.txt     # Crawler rules
│   └── llms.txt       # AI agent discovery
├── scripts/           # Build scripts
└── src/
    ├── components/    # React components
    ├── context/       # Theme context
    ├── pages/         # Page components
    └── styles/        # Global CSS

Scripts Reference

Script	Description
npm run dev	Start Vite dev server
npm run dev:convex	Start Convex dev backend
npm run sync	Sync posts to dev deployment
npm run sync:prod	Sync posts to production deployment
npm run sync:discovery	Update discovery files (development)
npm run sync:discovery:prod	Update discovery files (production)
npm run sync:all	Sync content + discovery (development)
npm run sync:all:prod	Sync content + discovery (production)
npm run import	Import URL as local markdown draft (then sync)
npm run build	Build for production
npm run deploy	Sync + build (for manual deploys)
npm run deploy:prod	Deploy Convex functions + sync to production

Tech Stack

• React 18
• TypeScript
• Vite
• Convex
• react-markdown
• react-syntax-highlighter
• date-fns
• lucide-react
• @phosphor-icons/react
• Netlify

Search

Press Command+K (Mac) or Ctrl+K (Windows/Linux) to open the search modal. The search uses Convex full text search to find posts and pages by title and content.

Features:

• Real-time results as you type
• Keyboard navigation (arrow keys, Enter, Escape)
• Result snippets with context around matches
• Distinguishes between posts and pages
• Works with all four themes

The search icon appears in the top navigation bar next to the theme toggle.

Real-time Stats

The /stats page shows real-time analytics powered by Convex:

• Active visitors: Current visitors on the site with per-page breakdown
• Total page views: All-time view count
• Unique visitors: Based on anonymous session IDs
• Views by page: List of all pages sorted by view count

Stats update automatically via Convex subscriptions. No page refresh needed.

How it works:

• Page views are recorded as event records (not counters) to avoid write conflicts
• Active sessions use heartbeat presence (30s interval, 2min timeout)
• A cron job cleans up stale sessions every 5 minutes
• No PII stored (only anonymous session UUIDs)

API Endpoints

Endpoint	Description
/stats	Real-time site analytics
/rss.xml	RSS feed with post descriptions
/rss-full.xml	RSS feed with full post content
/sitemap.xml	Dynamic XML sitemap
/api/posts	JSON list of all posts
/api/post?slug=xxx	Single post as JSON
/api/post?slug=xxx&format=md	Single post as markdown
/api/export	Batch export all posts with content
/meta/post?slug=xxx	Open Graph HTML for crawlers
/.well-known/ai-plugin.json	AI plugin manifest
/openapi.yaml	OpenAPI 3.0 specification
/llms.txt	AI agent discovery

Import External Content

Use Firecrawl to import articles from external URLs as markdown posts:

npm run import https://example.com/article

This will:

1. Scrape the URL using Firecrawl API
2. Convert to clean markdown
3. Create a draft post in content/blog/ locally
4. Add frontmatter with title, description, and today's date

Setup:

1. Get an API key from firecrawl.dev
2. Add to .env.local:

FIRECRAWL_API_KEY=fc-your-api-key

Why no npm run import:prod? The import command only creates local markdown files. It does not interact with Convex. After importing, sync to your target environment:

• npm run sync for development
• npm run sync:prod for production
• Use npm run sync:all or npm run sync:all:prod to sync content and update discovery files together

Imported posts are created as drafts (published: false). Review, edit, set published: true, then sync.

Theme Configuration

The default theme is Tan. Users can cycle through themes using the toggle:

• Dark (Moon icon)
• Light (Sun icon)
• Tan (Half icon) - default
• Cloud (Cloud icon)

To change the default theme, edit src/context/ThemeContext.tsx:

const DEFAULT_THEME: Theme = "tan"; // Change to "dark", "light", or "cloud"

Font Configuration

The blog uses a serif font (New York) by default. To switch fonts, edit src/styles/global.css:

Font Sizes

All font sizes use CSS variables defined in :root. Customize sizes by editing the variables:

Write Page

A public markdown writing page at /write (not linked in navigation).

Access directly at yourdomain.com/write. Content is stored in localStorage only (not synced to database). Use it to draft posts, then copy the content to a markdown file in content/blog/ or content/pages/ and run npm run sync.

Features:

• Three-column Cursor docs-style layout
• Content type selector (Blog Post or Page) with dynamic frontmatter templates
• Frontmatter field reference with individual copy buttons
• Font switcher (Serif/Sans-serif/Monospace)
• Theme toggle matching site themes
• Word, line, and character counts
• localStorage persistence for content, type, and font preference

AI Agent mode: When siteConfig.aiChat.enabledOnWritePage is enabled, a toggle button appears in the Actions section. Clicking it replaces the textarea with the AI Agent chat interface. The page title changes to "Agent" when in chat mode. Requires ANTHROPIC_API_KEY environment variable in Convex.

AI Agent Chat

The site includes an AI writing assistant (Agent) powered by Anthropic Claude API. Agent can be enabled in two places:

1. Write page (/write): Enable via siteConfig.aiChat.enabledOnWritePage. Toggle replaces textarea with Agent chat interface.

2. Right sidebar on posts/pages: Enable via aiChat: true frontmatter field (requires rightSidebar: true and siteConfig.aiChat.enabledOnContent: true).

Environment variables required:

• ANTHROPIC_API_KEY (required): Your Anthropic API key
• CLAUDE_PROMPT_STYLE, CLAUDE_PROMPT_COMMUNITY, CLAUDE_PROMPT_RULES (optional): Split system prompts
• CLAUDE_SYSTEM_PROMPT (optional): Single system prompt fallback

Set these in Convex Dashboard > Settings > Environment Variables.

Features:

• Per-page chat history stored in Convex (per-session, per-context)
• Page content can be provided as context for AI responses
• Markdown rendering for AI responses with copy functionality
• User-friendly error messages when API key is not configured
• Anonymous session authentication using localStorage
• Chat history limited to last 20 messages for efficiency
• System prompt configurable via environment variables (split or single prompt)

Source

Fork this project: github.com/waynesutton/markdown-site

License

This project is licensed under the MIT License.