How to Connect LLMtoMD to Cursor (MCP Setup in 60 Seconds)

Cursor is fast. Cursor's AI is great. But like every coding agent, it forgets your requirements as the project grows — and you end up re-pasting the spec. Connecting LLMtoMD gives Cursor a persistent, searchable memory of your FRD, API docs, and design decisions, retrieved on demand over MCP.

Here's the 60-second setup.

What you need

• Cursor v0.48+ (MCP support).
• A free LLMtoMD account with at least one document uploaded (your FRD, spec, or API reference). Connecting is free on every plan.

Option A — one click

On the LLMtoMD integrations page, click Add to Cursor. Cursor opens, asks you to confirm the install, and that's it — no file editing.

Option B — manual config

Create ~/.cursor/mcp.json (global) or .cursor/mcp.json (per-project) and add:

{
  "mcpServers": {
    "llmtomd": { "url": "https://mcp.llmtomd.com/mcp" }
  }
}

Sign in (first use)

The first time Cursor uses the server, it opens a browser to sign in to LLMtoMD — just approve it. No API key to manage; the connection is yours and you can revoke it anytime from Settings → Connect your AI tools in LLMtoMD.

Verify it works

In Cursor's chat, ask:

List my LLMtoMD documents.

You should see your library. Now the real payoff — point it at your spec:

Using my LLMtoMD documents, what are the auth requirements for this project?

Cursor pulls the answer straight from your FRD instead of guessing.

Why bother

Once connected, Cursor's agent can search_documents and ask_documents across your project knowledge base while it codes — so generated features stay aligned to the spec, and you stop re-explaining the same context every session. (We dug into the why, and the token savings, in Give Your AI Coding Agent a Memory.)

---

Make your project docs agent-ready. Convert your first document free →, then connect it to Cursor in a minute.