Filesystem
Provides filesystem operations through the Model Context Protocol, enabling read/write access and directory management.
Every MCP server in this guide was installed and tested in a real Cursor IDE environment. Here's our testing methodology:
Provides filesystem operations through the Model Context Protocol, enabling read/write access and directory management.
Integrates GitHub MCP Server as a context server for Zed's Assistant.
Community MCP server enabling AI agents to interact with Bluesky (AT-Protocol).
Official MCP server for GitHub integration, enabling AI assistants to manage repositories, issues, pull requests, and co...
An MCP server that integrates Brave Search API with AI assistants for web, image, video, news, and local search.
Fetches up-to-date documentation and code examples for libraries directly into LLM prompts.
Integrates powerful web scraping and content extraction capabilities into LLM clients like Cursor and Claude.
7 servers
5 servers
5 servers
2 servers
2 servers
2 servers
7 servers
Filesystem + Git + Fetch + GitHub + GitHub Actions + Context7 + Firecrawl
Add: Postgres/Supabase + Sentry/Grafana + Stripe + Notion/Jira
Add: Kubernetes + Grafana + Cloudflare (if using edge)
Add: Exa + Pinecone + DeepWiki + Vercel MCP Adapter
Does it solve a real problem in your daily dev loop?
Is it actively maintained with recent commits and updates?
Does it use scoped tokens and least privilege access?
Does it work with your tech stack (GitHub, Next.js, Supabase, etc.)?
Node.js 18+, Cursor 0.40+, npm/uv depending on server
Edit ~/.cursor/mcp.json or project .cursor/mcp.json
Add server command, args, and environment variables
Restart Cursor completely and test with @ mention
{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "."]
},
"git": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-git"]
},
"github": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-github"],
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "your_token_here"
}
}
}
}{
"mcpServers": {
"postgres": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-postgres"],
"env": {
"POSTGRES_CONNECTION_STRING": "postgresql://user:pass@localhost:5432/db"
}
},
"supabase": {
"command": "npx",
"args": ["-y", "@supabase/mcp-server"],
"env": {
"SUPABASE_URL": "https://your-project.supabase.co",
"SUPABASE_SERVICE_ROLE_KEY": "your_key_here"
}
}
}
}{
"mcpServers": {
"exa": {
"command": "npx",
"args": ["-y", "exa-mcp-server"],
"env": {
"EXA_API_KEY": "your_exa_key_here"
}
},
"context7": {
"command": "npx",
"args": ["-y", "@upstash/context7-mcp"]
}
}
}For detailed setup instructions, see our complete Cursor MCP setup guide or browse all MCP servers in our directory.
MCP servers for Cursor are integrations that connect Cursor to external tools (GitHub, databases, docs, CI/CD, observability, browsers) so Cursor can retrieve context and perform actions inside your workflow.
The best MCP servers in 2026 depend on your stack, but most teams start with: Filesystem, Git, Fetch, GitHub, GitHub Actions, Context7, and Firecrawl—then add Postgres/Supabase and Sentry/Grafana.
If you ship backend features or debug production issues, yes. A database MCP server helps you validate assumptions, inspect schemas, and generate safer SQL changes.
They can be safe if you use least privilege (scoped tokens), limit filesystem access, prefer read-only modes, and keep an inventory of installed servers.
Playwright MCP and Chrome DevTools MCP are excellent for reproducing UI bugs and validating changes, especially when paired with GitHub + Actions for CI.