Connect your AI agents to the world via Model Context Protocol
12 MCP servers found
The official collection of Model Context Protocol reference servers, maintained by Anthropic and the community. Includes Filesystem, Git, Memory, Fetch and more — the canonical starting point for learning and building MCP servers.
GitHub's official MCP server. Lets AI agents manage repositories, issues, pull requests and GitHub Actions directly, with both remote-hosted and local modes.
Microsoft's official Playwright MCP server. Drives browser automation via the accessibility tree instead of screenshots, letting agents browse pages, fill forms, capture screenshots and run tests without a vision model.
The official MCP server from the Chrome DevTools team. Lets agents debug a live Chrome browser: performance tracing, network request analysis, console debugging and page automation in one.
A real-time documentation MCP server by Upstash. Injects up-to-date, version-specific official library docs and code examples into coding agents, eliminating outdated-API hallucinations.
Cloudflare's official suite of remote MCP servers. Covers docs search, Workers builds, observability, browser rendering and a dozen more servers connecting agents to the Cloudflare platform.
Supabase's official MCP server. Lets agents manage Supabase projects directly: inspect tables, run SQL, apply migrations and manage config — the full database development loop.
Notion's official MCP server. Agents can read and write Notion pages and databases, search the workspace and manage comments — turning your knowledge base into an AI-operable workspace.
Firecrawl's official MCP server. Gives agents powerful web scraping capabilities: JS rendering, full-site crawling, search and structured data extraction with clean Markdown output.
An open-source coding agent toolkit. Provides symbol-level code retrieval and editing via language servers (LSP), giving any agent IDE-grade code understanding as an MCP server.
An MCP server connecting Blender to AI agents. Drive 3D modeling with natural language: create scenes, edit materials and run Blender Python scripts. A wildly popular community project.
An MCP server that provides Figma layout data to coding agents. Lets tools like Cursor implement designs in one shot — far more accurate than pasting screenshots.