The tool shed for AI agents.

Open-source personal AI agent that runs on your own machine and acts across chat apps, browser and OS.

Agentic skills framework and dev methodology: brainstorming, planning, TDD and subagent-driven development.

Nous Research's open-source agent that grows with you: persistent memory, skills and a gateway for chat platforms.

Open-source coding agent with a polished terminal UI that works with dozens of model providers.

Microsoft's converter from PDF, Office files, HTML and more into LLM-friendly Markdown, with an MCP server package.

Anthropic's public repository of Agent Skills, folders of instructions and scripts agents load on demand.

Anthropic's agentic coding tool that lives in your terminal and IDE, understands your codebase and executes multi-step changes.

GitHub's toolkit for spec-driven development with coding agents.

OpenAI's open-source, lightweight coding agent that runs locally in your terminal, with IDE and cloud counterparts.

Open-source library and cloud for AI browser agents that navigate and act on websites.

Google's open-source agent that brings Gemini models into the terminal, with MCP support and built-in tools.

The largest community-curated list of MCP servers, organized by category.

High-performance open-source editor with built-in agent panel and support for external agents via the Agent Client Protocol.

Open platform for AI software development agents that can write code, run commands and browse the web.

Open-source autonomous coding agent available as an IDE extension, CLI and SDK, with human-in-the-loop approvals.

Memory layer for AI agents and apps; includes OpenMemory for sharing memory across MCP clients.

Pulls up-to-date, version-specific library documentation and code examples into the prompt so agents stop hallucinating stale APIs.

AI gateway and SDK that calls 100+ LLM APIs in OpenAI format, with cost tracking and guardrails.

Framework for orchestrating role-playing, collaborating teams of autonomous agents.

Open-source, extensible on-machine agent (desktop + CLI) that uses MCP extensions and any LLM.

Agile AI-driven development method with specialized agent personas for planning through delivery.

Lets coding agents control and inspect a live Chrome: network, console, performance traces, Lighthouse audits and heap snapshots.

Framework for document-centric agents and RAG, with parsing and indexing tools.

Pioneering open-source AI pair programmer for the terminal that edits code and commits to Git.

Low-level orchestration framework for building resilient, stateful agents as graphs.

Framework and runtime for building, running and managing multi-agent systems.

Large multi-harness marketplace of subagents, skills and plugins for Claude Code, Codex, Cursor and more.

Microsoft's browser-automation server built on Playwright. Drives pages through accessibility snapshots instead of screenshots, so it is fast and deterministic.

Open-source coding agent for the IDE, terminal and CI, with configurable models and rules.

Open-source LLM engineering platform for tracing, evals, prompt management and metrics.

GitHub's official server for working with repos, issues, pull requests, Actions and code security from inside an agent. Available locally or as a hosted remote server.

Framework for real-time, temporally-aware knowledge graphs for agents, with a bundled MCP server for agent memory.

Community plugin connecting Blender to LLMs so agents can create, modify and render 3D scenes.

Integration platform with 1,000+ toolkits, managed auth and tool search for agents.

Coding toolkit that gives agents IDE-like semantic retrieval and symbol-level editing via language servers.

Lightweight framework from OpenAI for multi-agent workflows with handoffs, guardrails, tracing and MCP tools.

Hugging Face's barebones library for agents that think and act by writing code.

TypeScript framework for AI agents and apps: agents, workflows, memory, RAG, evals and MCP.

Charm's glamorous terminal coding agent with multi-model support, LSP integration and MCP.

Open-source terminal coding agent from the Qwen team, optimized for Qwen-Coder models.

Kanban board for orchestrating multiple coding agents in parallel, each in its own worktree.

AI task-management system that turns a PRD into dependency-aware tasks your coding agent works through.

The Pythonic way to build MCP servers and clients, with auth, proxies, OpenAPI generation and deployment helpers.

Open-source agentic engineering platform for VS Code, JetBrains and the CLI with hundreds of models.

Vercel's open-source TypeScript toolkit for building AI apps and agents across model providers.

Browserbase's SDK mixing natural-language actions with code to automate and extract from any site.

Platform for stateful agents with advanced long-term memory (from the MemGPT team).

Open, simple format, a README for agents, that tells coding agents how to build, test and work in your repo.

Official Python SDK for building MCP servers and clients.

Open-source VS Code agent with customizable modes (architect, code, debug) that act like a dev team.

Gives agents deep knowledge of n8n's nodes so they can design, validate and deploy n8n workflows for you.

Google's code-first toolkit for building, evaluating and deploying agents.

Type-safe Python agent framework from the Pydantic team, model-agnostic with MCP support.

CLI that analyzes local Claude Code (and other agent) usage logs to report tokens and costs.

Google's open-source server that exposes many databases (Postgres, MySQL, BigQuery, Spanner, AlloyDB and more) as safe, configurable tools.

Feeds simplified Figma layout and styling data to coding agents so they can implement designs more accurately.

Bridge between agents and the Unity Editor for managing assets, scenes, scripts and editor automation.

Secure cloud sandboxes for running agent-generated code with real tools.

Microsoft's framework for building and orchestrating agents and multi-agent workflows in Python and .NET.

Official TypeScript SDK for building MCP servers and clients.

Chrome extension-based server that exposes your everyday, logged-in browser (tabs, history, bookmarks, page content) to agents.

Exposes existing FastAPI endpoints as MCP tools, with auth, in a few lines of code.

Formerly Zen MCP. Lets Claude Code, Gemini CLI or Codex consult other models (Gemini, OpenAI, Grok, Ollama...) for reviews, debugging and consensus.

Exposes thousands of Apify Actors so agents can scrape social media, maps, e-commerce and search results.

Official visual tool for testing and debugging MCP servers: list tools, call them, inspect resources and prompts.

Full-stack framework for building MCP servers, MCP Apps for ChatGPT/Claude, and agents that connect to any MCP server.

Gives agents terminal control, process management, file search and diff-based editing on your machine.

AWS Labs' suite of specialized servers for AWS docs, CDK, cost analysis, serverless, databases and more.

Free remote server that turns any GitHub repo into a documentation source for agents, cutting code hallucinations.

Build custom agents on the same harness that powers Claude Code, with tools, hooks and MCP.

Official Firecrawl server for scraping, crawling, searching and extracting structured data from websites.

Community-driven, official registry and API for publishing and discovering MCP servers.

Popular community server for Jira and Confluence that supports both Cloud and Server/Data Center deployments.

Generates and searches modern React/Tailwind UI components from inside Cursor, Claude Code or Windsurf.

Exa's server for neural web search, code search and crawling, available locally or as a hosted endpoint.

Notion's official server for searching, reading and editing pages and databases from an agent.

Cloudflare's collection of remote MCP servers covering Workers, observability, DNS analytics, Radar, browser rendering and docs.

Catalog of Microsoft's official MCP servers, including Azure, Microsoft Learn docs and other first-party integrations.

Grafana's server for searching dashboards, querying Prometheus and Loki, and managing incidents and alerts.

Lets agents control cloud browsers on Browserbase, using Stagehand for natural-language actions and extraction.

Comprehensive server for Gmail, Calendar, Drive, Docs, Sheets, Slides, Forms, Tasks and Chat with OAuth 2.1.

Local server for searching arXiv, downloading papers and reading them section by section for literature work.

Official server for managing Supabase projects: run SQL, apply migrations, inspect logs, deploy edge functions and manage branches.

Tavily's production server for real-time search, page extraction, site mapping and crawling.

Figma's official server that gives agents design context, variables, components and Code Connect mappings from Figma files.

Stripe's official server and agent toolkit for creating customers, products, payment links and invoices, and searching Stripe docs.

Feature-rich Slack server for reading channels, DMs and threads and searching history, with multiple auth modes.

HashiCorp's server that gives agents live Terraform Registry data on providers and modules plus HCP Terraform workspace operations.

Docker's CLI plugin and gateway that runs catalog MCP servers in isolated containers and exposes them through one endpoint.

Brave's official server for web, news, image, video and local search via the Brave Search API.

Atlassian's official hosted server connecting Jira, Confluence, Compass and more to agents with OAuth.

Official server to query MongoDB databases and manage Atlas clusters with natural language.

Sentry's remote MCP server for pulling issues, stack traces and performance data into an agent so it can debug production errors.

Elastic's server for listing indices, inspecting mappings and running queries against Elasticsearch.

Neon's server for managing serverless Postgres projects and branches and running SQL through natural language.

Official Redis server for managing and searching data structures, including vector search, from an agent.

Reference server from the MCP maintainers that gives an agent read/write access to directories you explicitly allow, with configurable roots.

Reference server for reading, searching and manipulating local Git repositories: status, diffs, logs, commits and branches.

Reference server that fetches a URL and converts the page to clean Markdown so an agent can read web content efficiently.

Reference server providing persistent memory as a local knowledge graph of entities, relations and observations.

Reference server that gives a model a structured scratchpad for step-by-step, revisable reasoning on complex problems.

Kitchen-sink reference server exercising prompts, resources, tools and sampling. Useful for testing MCP clients.

PostHog's official server for querying product analytics, managing feature flags and investigating errors from your editor.

Vercel's official remote server for searching docs and managing projects, deployments and logs with OAuth.

Linear's official remote server for finding, creating and updating issues, projects and comments.

Hosted server that exposes actions across thousands of Zapier-connected apps to any MCP client.

Built into the shadcn CLI: lets agents browse, search and install components from shadcn/ui and any compatible registry.

Registry and hosting platform for finding, installing and running MCP servers and agent skills.

Community directory of MCP servers and clients with search and categories.

Searchable directory of open-source MCP servers with quality signals, plus hosted connectors.

AI-first code editor with agent mode, background agents and deep MCP support.

Agentic IDE built around the Cascade agent, with MCP support and codebase-aware flows.

GitHub's AI pair programmer with an agent mode in the IDE and an asynchronous coding agent that opens pull requests.

Cognition's autonomous AI software engineer that works in its own cloud environment and delivers PRs.

Agentic coding tool for the terminal and editors, designed to get the most out of frontier models.

Agentic development environment: a modern terminal with built-in coding agents and MCP support.

Platform of 'Droids', agents for coding, review and migrations that run in the terminal, IDE or CI.

Google's asynchronous coding agent that clones your repo into a cloud VM, makes changes and opens PRs.

Agentic IDE focused on spec-driven development: requirements, design and tasks before code.

Google's agent-first development platform for orchestrating multiple coding agents across editor, terminal and browser.