Skip to content
Field Academy
Create account

Agents & Workflows

How to Set Up MCP in Cursor

By The Field Academy Editorial TeamUpdated

MCP (Model Context Protocol) lets Cursor's agent call external tools and pull in context from outside your repo — Jira, Postgres, docs, and more. You add a server in Cursor's MCP settings (or a .cursor/mcp.json file), give it a command or URL, and the agent can then invoke its tools. Most setup problems are a bad command/path or too many tools enabled.

What is MCP in Cursor?

MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. is an open standard for connecting AI agents to external tools and data sources. In Cursor, an MCP server exposes tools (query a database, read a ticket, call an API) that the agent can use mid-task — so it isn't limited to your codebase.

How do I add an MCP server to Cursor?

  1. 1Open Cursor Settings → MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. (or create a .cursor/mcp.json in your project).
  2. 2Add the server: a command (stdio, e.g. npx -y some-mcp-server) or a remote URL (HTTP/SSE).
  3. 3Provide any required env vars / API keys for that server.
  4. 4Reload; confirm the server shows its tools as enabled (green).
  5. 5Ask the agent to use a tool — it will request approval before calling it.
// .cursor/mcp.json
{
  "mcpServers": {
    "example": {
      "command": "npx",
      "args": ["-y", "@example/mcp-server"],
      "env": { "API_KEY": "..." }
    }
  }
}

Why does my MCP server show 'no tools found'?

  • The command/path is wrong or the package isn't installed — test it in a terminal first.
  • Missing env vars / API keys so the server fails to start.
  • stdio vs remote mismatch — a remote server needs a URL, not a command.
  • Too many tools enabled across servers — trim to what you need so the agent isn't overwhelmed.

Frequently asked questions

What's the difference between stdio and remote MCP servers?

A stdio server runs locally as a command Cursor launches; a remote server is hosted and connected over HTTP/SSE with a URL. Use stdio for local tools, remote for shared/hosted ones.

Why isn't my MCP server connecting in Cursor?

Most often the start command fails — wrong path, missing dependency, or a missing API key. Run the exact command in a terminal to confirm it starts, then check the env vars in your MCP config.

Can I limit which MCP tools the agent uses?

Yes — enable only the servers and tools you need. Fewer, relevant tools produce more reliable agent behavior than a large, noisy tool list.

Sources & last verified

Cursor ships frequently. Facts verified against primary sources on June 15, 2026.

Keep reading