How to add an MCP server to Cursor

Cursor's agent uses MCP servers for anything beyond the codebase: issue trackers, docs, databases, browsers, design files. This guide covers adding them, the config format, sharing across a team, and the enterprise setup. For what to add, see best MCP servers for teams.

Adding a server through the UI

1. Open Cursor Settings (Cmd/Ctrl + Shift + J).
2. Go to Tools and Integrations, then MCP.
3. Click Add new MCP server. Cursor opens mcp.json for you.
4. Add the server (format below), save, and toggle it on. The tools list appears under the server once it connects.

The config format

Cursor reads an mcp.json with an mcpServers object. Local servers use command and args; remote servers use url.

{
  "mcpServers": {
    "linear": {
      "url": "https://mcp.linear.app/mcp"
    },
    "playwright": {
      "command": "npx",
      "args": ["-y", "@playwright/mcp@latest"]
    },
    "postgres": {
      "command": "npx",
      "args": ["-y", "@bytebase/dbhub", "--dsn", "${env:DB_DSN}"],
      "env": { "NODE_ENV": "production" }
    }
  }
}

Environment variables go in env, or are referenced with ${env:NAME} so secrets stay out of the file.

Project vs global

• Global: ~/.cursor/mcp.json. Your servers, every project.
• Project: .cursor/mcp.json in the repository root. Commit it and the whole team gets the same servers.

The project file is the right place for the team's shared tools. Keep credentials out of it with environment references.

One-click installs

Cursor has a directory of MCP servers and supports install links. Clicking Add to Cursor on a vendor's page writes the entry into your mcp.json. Convenient, and exactly the moment to read what is being written: a link can add any command. Vendor pages are fine; random blog posts less so.

Authentication

Remote servers that need login trigger an OAuth flow when you enable them; Cursor stores the token. If a server shows an error after a while, disable and re-enable it to re-authenticate.

Using tools in the agent

Once a server is connected, the agent sees its tools and proposes calls when relevant. You approve each call by default; you can enable auto-run for tools you trust. Tool results land in the conversation as context, which is why servers that return content others control (tickets, web pages, docs) deserve the caution described in MCP security best practices.

Standardising across a team

Cursor's project mcp.json gets everyone the same server list, but not the same credentials or the same policy, and it does nothing for the people using Claude Code or Copilot alongside Cursor.

For a company, the pattern is a single gateway server in mcp.json that provides every approved server behind per-user policy, central credentials and one log. Walma AI Hub does that for Cursor, Claude Code and Codex, running inside the customer's own EU region. See What is an MCP gateway?.