Home

waynesutton/markdown-site

Status:Indexed
Analyzed Dec 22, 2025300+ starsTypeScript

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

Architecture Overview

This library provides a framework for dynamically publishing content from Markdown files to a real-time web experience. Developers interact with this system primarily through configuration files and by authoring Markdown content. The framework's core value proposition is its ability to synchronize content via Convex, making it accessible to browsers, LLMs, and AI agents without traditional build and deploy cycles. This real-time synchronization is a key feature, enabling live content updates and a seamless user experience.

Internally, the framework is built around a real-time synchronization engine and a flexible content processing pipeline. The Content Synchronization Service is powered by Convex and handles the real-time updates of markdown posts and pages, ensuring that local changes are instantly propagated to all connected clients. This service is crucial for the framework's "no rebuild/redeploy" publishing model. Under the hood, the Content Parsing and Processing Engine takes raw Markdown content and its frontmatter, transforming it into a structured JavaScript object. This engine, along with a separate Markdown Parser component, ensures that all content is consistently processed into a usable format, making it machine-readable and ready for manipulation by other systems.

To enhance discoverability, the framework exposes a comprehensive set of public-facing endpoints and files designed for AI agents and LLMs. The AI and LLM Discovery APIs package provides OpenAPI specifications, AI plugin manifests, and various API endpoints for listing and retrieving posts. This meticulous approach to discoverability is a key differentiator, allowing AI agents to programmatically interact with and utilize the site's content. Complementing this is the SEO and Metadata Generation package, which dynamically creates essential metadata like sitemaps, RSS feeds, and structured data, further improving content visibility for both search engines and AI crawlers.

What makes this library special is its commitment to a real-time, serverless content publishing model. By leveraging Convex, it eliminates the need for traditional build processes, allowing for instant updates. This architectural choice is a trade-off that prioritizes immediate content availability and developer velocity over the predictable, but often slower, nature of static site generation. For developers looking to extend the framework, understanding the flow of content from Markdown files through the parsing engine, into the Convex database via the Real-time Data Reconciliation Service, and finally being rendered by the Content Rendering and Theming components is paramount. The Core Configuration and Site Setup package is also a vital entry point for customization.

Architecture Diagram

Rendering diagram...

Data Flow

Rendering diagram...
Analysis completed in 3 iterations • Discovered 5 packages, 2 modules, 1 components