MCP clientsAI

Connect OpenClaw to thirds.ai

Use with your agent

Give this prompt to your AI agent. The agent does the task for you.

Read https://thirds.ai/docs/mcp/openclaw and help me connect an OpenClaw runtime to thirds.ai and verify its tools before scheduling file creation

Use the embedded OpenClaw runtime with its coding or messaging tool profile. Other runtime adapters can support different transports. See the OpenClaw poster workflow for a weekly example.

Add the server

Create a key on the API keys page. Add this entry to your private OpenClaw configuration. Replace YOUR_API_KEY locally:

JSON
{
  "mcp": {
    "servers": {
      "thirds": {
        "url": "https://thirds.ai/mcp",
        "transport": "streamable-http",
        "headers": {
          "Authorization": "Bearer YOUR_API_KEY"
        }
      }
    }
  }
}

Keep this file private and out of source control. Do not enable OAuth: thirds.ai uses the bearer key. Read OpenClaw's transport settings for credential handling.

Check the running agent

Run openclaw mcp probe thirds to check the live connection. Then restart the Gateway or agent process and ask it to list the thirds.ai tools. Saving an entry alone does not prove the active runtime can use it. The minimal profile hides MCP tools.

The OpenClaw registry guide explains runtime support and reloads.

Schedule the work

Run one approved example first. Then configure the schedule and input source in OpenClaw. thirds.ai makes the files; OpenClaw starts each run. Read costs and retries before enabling unattended work.

Ready to make your first file?

Start in the playground with a gallery template, or get an API key and send your first request from your own terminal.