MCP Servers
The open registry for Model Context Protocol servers. Find the right tools, resources, and prompts for your AI agents — filtered by category, transport, or use case.
Servers
92
Tools
343
Categories
11
Contributors
79
Provides comprehensive search capabilities through the Brave Search API via the Model Context Protocol. Enables AI agents to perform web searches, local business searches, image searches, video searches, news searches, and AI-powered summarization. Supports both STDIO and HTTP transports.
Official Elastic MCP server that connects AI agents to Elasticsearch data using the Model Context Protocol. Enables natural language interactions with Elasticsearch indices — querying, analyzing, and retrieving data without custom APIs. Supports both stdio and streamable-HTTP transports, and works with Elasticsearch 8.x/9.x clusters including Elasticsearch Serverless. Distributed as a Docker container image from the Elastic registry.
Official MCP server for the Perplexity API Platform. Provides AI agents with real-time web search, deep research, and advanced reasoning capabilities through Sonar models. Includes tools for quick web search, conversational Q&A with citations, comprehensive deep research reports, and complex analytical reasoning. Returns answers with source attribution and supports configurable timeouts for long research queries.
Official Expo MCP server that connects AI coding assistants to Expo projects and EAS services. Enables searching and reading Expo documentation, managing EAS builds and workflows, installing compatible libraries, inspecting TestFlight crashes and feedback, and automating visual verification through simulator screenshots and interactions. Supports both remote server capabilities and local development server features for advanced automation.
GitHub's official MCP Server that connects AI tools directly to GitHub's platform. Enables AI agents to manage repositories, issues, pull requests, branches, files, actions workflows, and code security. Supports both remote (OAuth) and local (Docker/binary) modes with fine-grained toolset configuration.
Official Vercel MCP server that gives AI tools secure access to Vercel projects via OAuth. Enables searching Vercel documentation, managing projects and deployments, analyzing deployment logs, and interacting with Vercel infrastructure. Supports Streamable HTTP transport with OAuth authentication and integrates with Claude Code, Cursor, VS Code, ChatGPT, Codex CLI, and other AI assistants.
Official Figma MCP server that brings design context directly into AI coding workflows. Provides tools for extracting design information, generating code from Figma selections, taking screenshots, creating and editing Figma files, generating diagrams from Mermaid syntax, searching design systems, managing Code Connect mappings, and uploading assets. Supports both remote (OAuth) and local (desktop app) server modes.
MCP server for searching and accessing arXiv research papers. Enables AI agents to search papers with filters for date ranges and categories, download full paper content, read papers in markdown format, and perform semantic search across locally stored papers. Supports citation graph exploration via Semantic Scholar and research alert watches for tracking new publications on topics of interest.
Official Hugging Face MCP Server that connects AI assistants directly to the Hugging Face Hub ecosystem. Provides tools for searching and retrieving models, datasets, and research papers, running inference on thousands of Gradio-powered AI applications (Spaces), and accessing the full Hub API. Supports remote HTTP-streaming via https://huggingface.co/mcp with OAuth or Bearer token auth, as well as local stdio deployment. Works with Claude, Gemini CLI, VS Code, Cursor, and any MCP-compatible client.
Official Atlassian Rovo MCP Server — a cloud-based bridge between Atlassian Cloud and any MCP-compatible AI tool. Enables AI agents to search, summarize, create, and update Jira issues, Confluence pages, and Compass components in real-time. Uses OAuth 2.1 or API token authentication, respects existing user permissions, and supports remote HTTP-streaming as well as local stdio via the mcp-remote proxy. Works with Claude, GitHub Copilot, Gemini CLI, VS Code, Cursor, and ChatGPT.
Official Asana MCP server that connects AI tools to the Asana Work Graph. Enables agents to create and update tasks, manage projects and sections, add comments, search work, and summarize project status via a remote OAuth-secured endpoint.
Official HubSpot MCP server connecting AI tools to the HubSpot CRM. Enables agents to read and manage contacts, companies, deals, and tickets, search the CRM, and create engagements such as notes and tasks. Honors HubSpot scopes and rate limits via OAuth.
Official MCP server for Qdrant vector search engine. Acts as a semantic memory layer enabling AI agents to store and retrieve information using vector similarity search. Supports storing text with metadata, semantic querying, configurable embedding models via FastEmbed, and both cloud-hosted and local Qdrant instances. Useful for building RAG pipelines, code search, knowledge bases, and long-term agent memory.
Official Apify MCP server that gives AI agents access to thousands of Actors for web scraping and automation. Agents can run Actors to extract data from websites, search engines, social platforms, and maps, then retrieve structured results from the dataset, enabling rich web research and data collection.
Skills vs MCP servers
what's the difference?Skillsthe “what to do”
A skillA reusable, structured prompt/workflow with recommended models, an example prompt, and compatible tools. packages know-how — instructions, an example promptA ready-to-use prompt template that demonstrates how to invoke the skill., and recommended models — so an agent performs a task consistently. Skills add knowledge, not new connections.
MCP serversthe “how to connect”
An MCP serverModel Context Protocol server — a standard way to expose tools, resources, and prompts to AI agents and IDEs. gives an agent new capabilities by connecting it to real systems (databases, APIs, files) over a transportHow the client talks to the server: stdio (local process), SSE, or HTTP streaming.. MCP adds connections and actions, not task instructions.
Rule of thumb: reach for a skill when you need the model to do a task well, and an MCP server when you need it to reach a tool or system. They compose — a skill can rely on tools an MCP server provides.
Built an MCP server?
Submit it to the registry — it's open source and community-maintained.