Skip to content
Letos is in beta.What that means

Guides

Connect an AI assistant (MCP)

How to connect Claude Desktop or Claude Code to Letos over MCP in a few minutes, safely and read-only.

~8 min

Letos has its own MCP server in the desktop app. With it, an AI assistant such as Claude Desktop or Claude Code can be asked in plain words (“What is on my plate this week?”) and it reads your tasks, goals, calendar and mail in Letos. This guide connects one safely and read-only.

The server runs only on your computer (127.0.0.1). It can’t be reached from another computer or from the internet. There’s no MCP on the iPhone.

1. Turn on agents

  1. Open Settings → Agents.
  2. Turn on Allow agents (MCP). The first time you see the consent Send data off this device: what agents will see and that the agent may pass it on to its provider. The server starts only after Allow.

2. Add an agent (read-only to start)

  1. Under Agents with access click Add agent.
  2. Enter a name (for example “Claude Desktop”).
  3. Choose the Read permission. Start with read — writes are off by default.
  4. Copy the token and the client setup that appear. You will see them only once.

3. Connect the client

Claude Code — click Copy Claude Code command and run it in a terminal:

claude mcp add --transport http letos http://127.0.0.1:47800/mcp --header "Authorization: Bearer <agent token>"

Claude Desktop — click Copy Claude Desktop setup and put the block in claude_desktop_config.json (on macOS in ~/Library/Application Support/Claude/):

{
  "mcpServers": {
    "letos": {
      "command": "/Applications/Letos.app/Contents/MacOS/letos",
      "args": ["mcp"],
      "env": { "LETOS_MCP_TOKEN": "<agent token>" }
    }
  }
}

Treat the command and the token as secrets.

4. Try it

Letos must be running. Restart the client and try a simple prompt:

  • “List my tasks for today.”
  • “Which unread mail from the last three days needs an answer from me?”

5. Turn on writes when you want them

Only once you know what you need them for:

  1. Set the agent’s permission to Write or Read and write.
  2. Turn on the main Allow writes switch.

The agent can then create and edit tasks, add a step to a goal, create an event without guests, or save an e-mail draft. It deletes nothing, and an e-mail is sent only after you confirm it.

Watch out for injected instructions. An agent also reads texts written by others (e-mails, notes) that can contain hidden instructions. Turn writes on thoughtfully and connect only agents you trust. You see every call in Settings → Agents → Activity log.

What next

Last updated: