The directory of MCP servers

Provides secure file system access with configurable allowed directories. Supports reading, writing, listing, searching, and moving files. The official reference server for local file operations in Claude Desktop.

Exposes Git repository operations — reading commits, diffs, branches, and file contents — directly to your AI agent. Lets Claude understand and navigate codebases without spawning shell processes.

Fetches web pages and converts them to clean Markdown, making web content accessible to Claude without browser overhead. Handles redirects, respects robots.txt, and optionally returns raw HTML.

Gives AI agents persistent memory via a knowledge graph stored locally. Supports creating, reading, updating, and deleting entities and relations so Claude can remember facts across conversations.

Provides a structured tool for multi-step reasoning and problem decomposition. Claude uses it to break complex tasks into sequential thoughts, backtrack when needed, and produce well-reasoned final answers.

Provides current date and time in any timezone and converts between timezones. A simple, indispensable reference server that gives Claude accurate temporal context without hallucination.

A comprehensive test and demo server that implements every MCP feature: prompts, resources, tools, sampling, and logging. The definitive reference for MCP server authors learning the full protocol.

Full GitHub API integration — manages repos, issues, pull requests, code search, file reads, and branch operations. The official server for AI-driven GitHub workflows in Claude Desktop.

Integrates Claude with GitLab projects, merge requests, issues, pipelines, and repositories. Supports self-hosted GitLab instances as well as GitLab.com via personal access tokens.

Automates Chromium via Puppeteer so Claude can navigate pages, click elements, fill forms, take screenshots, and run JavaScript. Ideal for web scraping, UI testing, and RPA workflows.

Connects Claude to Brave Search's API for real-time web and local search results. Returns structured snippets, titles, and URLs without tracking — a privacy-respecting alternative to Google Search.

Provides geocoding, reverse geocoding, place search, directions, and distance matrix via the Google Maps Platform APIs. Lets Claude answer location questions and plan routes with real data.

Reads, lists, and searches files in Google Drive including Docs, Sheets, Slides, and PDFs. Converts Google Workspace documents to plain text for Claude to summarize and analyze.

Reads channels, threads, and messages from Slack workspaces. Lets Claude search conversations, summarize discussions, draft replies, and interact with your team's Slack data programmatically.

Queries Sentry issues, events, and stack traces so Claude can help diagnose production errors. Fetch recent exceptions, drill into specific error groups, and get AI-powered root cause analysis.

Interacts with the Stripe API to retrieve customers, payments, subscriptions, and refund data. Lets Claude answer billing questions and perform read operations on your Stripe account.

Connects Claude to a PostgreSQL database with read-only query execution and schema inspection. Ask natural language questions about your data and get back structured results without writing SQL manually.

Read and write SQLite databases from Claude — execute SQL, inspect schema, create tables, and insert data. Includes a built-in memo system that persists notes inside the database between sessions.

Queries Amazon Bedrock Knowledge Bases for RAG-powered retrieval. Fetches semantically relevant document chunks from your private knowledge base to ground Claude's responses in your own data.

Manages Cloudflare Workers, KV namespaces, D1 databases, R2 buckets, and DNS records via the Cloudflare API. Deploy code and manage edge infrastructure through natural language commands.

Reads and writes Notion pages, databases, and blocks. Claude can search your workspace, create pages from templates, update database properties, and draft rich content directly in Notion.

Connects Claude to Linear project management — query issues, cycles, projects, and teams. Create issues, update statuses, and summarize sprint progress using natural language.

Browser automation via Microsoft Playwright — navigate pages, interact with elements, take screenshots, and extract content across Chromium, Firefox, and WebKit. The most complete browser MCP for cross-browser testing.

Searches the web using Exa's neural search engine, which returns semantically relevant full-content results rather than just blue links. Great for research, fact-checking, and sourcing citations.

Crawls entire websites and converts them to clean Markdown, JSON, or structured data. Handles JavaScript-rendered pages, sitemaps, and multi-page crawls — the go-to MCP for deep web data extraction.

Reads, writes, and queries Redis key-value stores including strings, hashes, lists, sets, and sorted sets. Lets Claude inspect cache state, debug session data, and manage queue contents.

Queries and mutates MongoDB collections — find, insert, update, and delete documents with natural language. Inspect indexes, run aggregation pipelines, and explore your schema through Claude.

Connects Claude to MySQL and MariaDB databases for schema exploration and read-only SQL queries. Useful for data analysis, report generation, and database audits via natural language.

Manages Vercel deployments, projects, domains, and environment variables via the Vercel API. Trigger deploys, inspect build logs, and configure your Vercel infrastructure through Claude.

Reads and writes notes in an Obsidian vault — open files, create pages, search backlinks, and modify frontmatter. Turns Claude into a personal knowledge assistant for your second brain.

Queries and updates Atlassian Jira projects, issues, sprints, and boards. Claude can create tickets, update priorities, summarize backlogs, and search issues using JQL via natural language.

Reads, searches, drafts, labels, and threads Gmail messages via the Gmail API. Claude can compose emails from context, triage inboxes, and automate email workflows through OAuth-secured access.

Creates, reads, and updates Google Calendar events across multiple calendars. Let Claude schedule meetings, check availability, set reminders, and summarize your upcoming week.

Reads messages and channel history from Discord guilds via the Discord API. Claude can search conversations, summarize threads, and help moderate or draft content for your Discord community.

Controls Docker containers, images, volumes, and networks via the Docker socket. Claude can start or stop containers, inspect logs, pull images, and manage your local Docker environment.

Interacts with Kubernetes clusters via kubectl — lists pods, deployments, and services; reads logs; applies manifests; and executes commands inside containers. AI-driven cluster management.

Triggers workflows, reads run logs, and queries job statuses on GitHub Actions via the GitHub REST API. Let Claude investigate flaky CI runs and trigger deploys without leaving your chat.

Real-time web search and news retrieval powered by Tavily's AI-optimized search API. Returns clean, LLM-ready summaries and sources — built specifically for grounding AI agent responses.

Queries the Perplexity AI search API to get cited, up-to-date answers with source URLs. Adds real-time knowledge retrieval to Claude without leaving the MCP interface.

Reads Terraform state files and plans so Claude can explain infrastructure changes, review drift, and answer questions about your cloud resources. Supports local state and remote Terraform Cloud.

Reads and writes Airtable bases, tables, and records via the Airtable REST API. Claude can query filtered views, create records in bulk, and use Airtable as a lightweight app database.

Queries Datadog metrics, monitors, logs, and dashboards via the Datadog API. Claude can diagnose performance anomalies, review alert history, and correlate traces with logs in one conversation.