Mdreview is an MCP (Model Context Protocol) server designed for AI agents and developers. It renders Markdown files into strictly sanitized HTML, providing temporary, Tailscale-accessible preview URLs.

How does Mdreview ensure content security?

Mdreview features strict XSS protection using `bluemonday`, path validation to prevent directory traversal outside a configured workspace, and generates unguessable preview URLs with random UUIDs.

What is the benefit of Tailscale integration?

Tailscale integration automatically detects your Tailscale IP, ensuring that generated preview URLs are accessible from any device within your private Tailnet, simplifying secure sharing and review.

How can I integrate Mdreview with my AI development workflow?

Mdreview integrates seamlessly with Gemini CLI and Claude Code as an MCP server. AI agents can call its `preview_markdown` tool, providing a Markdown file path to receive a secure, shareable HTML preview URL.

Mdreview

Name: Mdreview
Author: sopranoworks

0•

Renders Markdown files to sanitized HTML for AI agents, providing Tailscale-accessible preview URLs.

Mdreview is a Model Context Protocol (MCP) server designed to enable AI agents to transform Markdown content into secure HTML. It facilitates rapid previewing of rendered Markdown across a Tailnet by automatically detecting the user's Tailscale IP and generating temporary, uniquely identified preview URLs. Built with a security-first approach, it includes robust XSS protection, path validation, and symlink resolution, ensuring a safe and efficient workflow for AI-driven content generation and review.

主な機能

01Markdown Rendering with high-fidelity HTML output

02Strict XSS Protection using bluemonday

03Tailscale Integration for network-wide accessible preview URLs

04Path Validation to prevent directory traversal

05In-memory preview store with unique UUIDs

060 GitHub stars

ユースケース

01Enabling AI agents (e.g., Gemini CLI, Claude Code) to preview Markdown-generated content.

02Securely sharing HTML previews of Markdown documents within a Tailscale network.

03Providing a controlled environment for AI to render and validate documentation or code explanations.