Connect

Connect Grok Build to Rafter’s MCP server

Grok Build takes remote MCP servers from its own config. One command and xAI's terminal agent can search your team's shared memory, follow its skills, and save what it learns back.

You need a Rafter account — create one if you haven’t.Takes about 2 minutes.No API keys — sign-in is OAuth, in the browser.
  1. Add the server

    Run this in your terminal.

    grok mcp add --transport http rafter https://app.heyrafter.xyz/api/mcp

    That writes the entry to ~/.grok/config.toml. A project-scoped .grok/config.toml works the same way.

  2. Or add it by hand

    The same thing, written straight into ~/.grok/config.toml if you would rather keep it in version control.

    [mcp_servers.rafter]
    url = "https://app.heyrafter.xyz/api/mcp"
  3. Sign in and pick a workspace

    Rafter needs a sign-in, so Grok opens a browser flow on the first call — Google, GitHub, or email — and you pick the workspace to connect. The token is kept in ~/.grok/mcp_credentials.json.

  4. Check it connected

    Type /mcps in the TUI to open the MCP tab and see Rafter with its tools. Then try “Search Rafter: how do we deploy?”

Your first five minutes

The connection is the easy part. The habit is what pays: make Grok Build check the shared brain before it improvises. Three things to try in Grok Build:

To make it automatic, paste this line into Grok Build’s project or system instructions:

Check Rafter first: search our skills and memories before you answer, and save anything durable back with add_memory.

A longer startup prompt and a ready-made Claude Code skill live in the rafter-mcp repo.

Troubleshooting

Is this the Grok chat app or the terminal agent?

The terminal agent — Grok Build, the grok binary. It is the one that reads MCP servers from a config file, which is what these steps set up.

Grok's tools have odd names

Grok namespaces tools from every MCP server as <server>__<tool>, so Rafter’s search arrives as rafter__search. That is expected, not a misconfiguration.

Which workspace should I pick when signing in?

The one whose skills and memories you want this tool to see. A connection sees exactly what you can see — your teams and your role — never more. You can connect again later and grant a different workspace.

Is my team's content used to train AI models?

No. Rafter does not use your content to train generalized or foundation models — see the privacy policy.

Give your whole team the same brain.From $29 per user — see pricing.

Create your account