What are the best MCP servers to connect to an AI agent?
Start with servers for the systems you already query by hand: your database, your error tracker, your issue tracker, your deploys. MCP is the standard that lets any compatible assistant call them without custom glue.
Surveyed 10 August 2026
Connect tools with MCP
Open in Tools →- 01Chrome DevTools MCP
MCP server exposing Chrome DevTools to agents — inspect, debug, profile pages programmatically.
AI Agents - 02Playwright MCP
Microsoft's MCP server for Playwright. Lets agents drive a browser via tool calls.
AI Agents - 03Context7
Up-to-date, version-specific library documentation for coding agents, served over MCP. Pulls current docs and working code examples into the prompt so the model stops writing against stale or hallucinated APIs.
Developer Tools - 04GitHub MCP Server
GitHub's official MCP server connects any coding agent to GitHub — repositories, issues, pull requests, Actions, code search, and security alerts — through the Model Context Protocol. The single most widely-used MCP server.
Developer Tools - 05Figma MCP Server
The Figma MCP Server connects Figma design files directly to AI coding agents via the Model Context Protocol, enabling them to extract design context, generate code from frames, and write native Figma content back to the canvas. It bridges the gap between design and development by giving AI agents access to variables, components, layout data, and design system resources. Ideal for product teams looking to accelerate design-to-code workflows.
Developer Tools - 06Notion MCP Server
An official Model Context Protocol (MCP) server for the Notion API, enabling AI agents and tools like Claude and Cursor to read, create, and edit Notion pages and databases. It supports both a local npm/Docker setup and a newer remote OAuth-based server, with optimized Markdown-based page editing for token efficiency.
Developer Tools - 07Slack MCP Server
A powerful Slack MCP server that connects an agent to Slack with no admin permissions required — read channels and DMs, post messages, and fetch smart history — supporting stdio and SSE transports, including GovSlack.
Developer Tools - 08Supabase MCP
Supabase MCP server — let your AI agent query, mutate, and manage Supabase projects through the Model Context Protocol.
Developer Tools - 09Sentry MCP
MCP server for Sentry. Lets Claude and other agents read issues, traces, and replays straight from your Sentry org.
AI Agents - 10Stripe MCP
Stripe's official MCP server: lets an agent search the Stripe docs, inspect account data, and call the Stripe API — create customers, products, payment links, invoices — with scoped keys.
Developer Tools - 11Cloudflare MCP Server
Cloudflare's official MCP servers: manage Workers, KV, R2, D1 and DNS, read analytics and logs, and search the Cloudflare docs from any MCP client.
Developer Tools - 12Vercel MCP
Vercel's official hosted MCP server: gives an agent access to projects, deployments, build logs and docs, so it can investigate failed builds and manage deployments without leaving the editor.
Developer Tools
Ordered by the VibeLeaderboard index, re-surveyed each edition. Something missing?
What to look for
- 01Are the permissions scoped and read-only where possible? An MCP server inherits whatever credentials you give it.
- 02How many tools does it expose? Servers with dozens of tools crowd the context window and degrade tool choice.
- 03Does the tool output stay small? Servers that return unbounded payloads exhaust the window in one call.
Common questions
- What is MCP?
- The Model Context Protocol, an open standard for exposing data and actions to AI assistants. Write one server and any compatible client can use it, instead of building a separate integration per assistant.
- What are the security risks of MCP servers?
- The server runs with the credentials you hand it, and data it returns can carry injected instructions. Scope permissions tightly, prefer read-only, and treat everything a server returns as untrusted content.
More in Work with agents
- Code with an agentWork across a repository from an issue, prompt, or terminal session.
- Control a browserLet an agent navigate, inspect, test, and complete workflows on the web.
- Research with an agentFind sources, verify claims, synthesize evidence, and preserve citations.
- Give an agent memoryPersist project context, decisions, preferences, and reusable knowledge.