Connect Paso to your AI

Add new tasks, re-plan them or analyze your work done in Paso using AI of your choice.

Concentric copper rings, an illustration of a connection

Paso’s MCP server lets an AI assistant work on your timeline the way you do. It can read what is planned, add tasks and notes, and move things between days and the backlog. It works on your account, so everything it adds shows up on all your devices, like anything you type yourself.

The server follows the Model Context Protocol (MCP), an open standard that Claude, ChatGPT, Cursor and many other AI clients support. Any of them connects with one address and your Paso login. No API keys. No configuration files.

It takes about a minute, and the steps below work the same in every client.

Before you start

  • Premium. The MCP server is part of Paso Premium. On the free plan, Settings > Integrations shows the plans.
  • A client that supports remote MCP servers. Most current AI clients do. If yours asks for a command instead of an address, see Clients without remote MCP support.
  • Your Paso login. The connection is tied to the account you sign in with, and to every project that account can see.

Setup

  1. Copy the address

    You will also find it in Paso under Settings > Integrations > MCP, ready to copy.

    https://mcp.paso.to

  2. Add it in your client

    Claude

    Open Claude app → Settings → Connectors → Add → Add custom connector

    ChatGPT

    Open ChatGPT → Settings → Plugins → Add → Add MCP server

  3. Approve access

    A browser window opens with the Paso sign-in page. Sign in with the account you use in the app and allow access. The window closes on its own.

  4. Try it

    Back in your client, ask: “What is on my timeline this week?” If the assistant answers with your tasks, you are connected.

How it works

Once connected, the assistant sees Paso as a set of tools it can call when your request needs them. You do not have to name the tools. You ask for what you want, in your own words.

Clients without remote MCP support

Some clients only run local servers and expect a command rather than an address. Use the mcp-remote bridge, which needs Node.js installed:

npx -y mcp-remote https://mcp.paso.to/

Or, as a configuration block, where your client keeps them:

{
  "mcpServers": {
    "paso": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.paso.to/"]
    }
  }
}

The bridge opens the same sign-in window and keeps the session for you.

Frequently asked questions

Yes. The MCP server is part of Paso Premium. On the free plan, Settings > Integrations shows the plans and what they include.

Any client that supports remote MCP servers. Today that includes Claude on the web, on desktop and in Claude Code, ChatGPT, Cursor, Visual Studio Code and many others, with new ones appearing every month. We do not keep step-by-step instructions for each client, because every client documents its own settings. Look for "custom connector", "add MCP server" or "custom server" in your client, and paste the address.

Everything you can see in Paso: your timeline and backlog, the projects you own and the projects you were invited to, including calendar events synced to them. It works under your account. Every change it makes syncs to all your devices and, in a shared project, other members see it as made by you.

Remove the Paso server from your client's settings. To revoke access on Paso's side as well, go to Settings > Integrations > MCP, find the client under Connected clients and remove it. The client will have to sign in again to reconnect.

Disconnect and reconnect the server in your client's settings. This resets the client side only. Nothing in Paso changes, and your data is not affected.

No. The server speaks Streamable HTTP only, which every current MCP client supports. Clients that still need SSE can use the mcp-remote bridge.

Paso does not use your data to train anything, and does not sell or share it. The MCP server sends your data only to the client you connected, and only when you ask that client to do something. What the client does with it is covered by the client's own privacy policy.