Make a meeting bot speak, from Python

What it does

The script waits for agent-ready, sends one say, and prints the server's agent-said confirmation before leaving. What it says comes from the command line.

It needs a key and a meeting code in MEETFUL_KEY and MEETING. Two curl requests make both. MEETFUL_URL is optional and defaults to https://meetful.com.

Python 3.9 or later. The socket scripts use the websockets package (pip install websockets; this ran with 13.1); the transcript script uses only the standard library.

The code

# Say something in a Meetful meeting from Python. pip install websockets
# MEETFUL_KEY=tmk_… MEETING=abc-defg-hij python speak.py "Noted: the launch moves to the third of March."
import asyncio, json, os, sys
import websockets

BASE = os.environ.get("MEETFUL_URL", "https://meetful.com")
TEXT = sys.argv[1] if len(sys.argv) > 1 else "Hello from Python."

async def main():
    async with websockets.connect(BASE.replace("http", "ws", 1)) as ws:
        await ws.send(json.dumps({"type": "agent-hello", "key": os.environ["MEETFUL_KEY"], "meeting": os.environ["MEETING"]}))
        async for raw in ws:
            m = json.loads(raw)
            if m["type"] == "agent-ready":
                await ws.send(json.dumps({"type": "say", "text": TEXT}))
            elif m["type"] == "agent-said":
                print(f"{m['by']} said: {m['text']}" + (" (read aloud)" if m.get("audio") else ""))
                return
            elif m["type"] in ("agent-error", "agent-refused"):
                raise SystemExit(f"{m['type']}: {m.get('why')}")

asyncio.run(main())

Run it

MEETFUL_KEY=tmk_… MEETING=abc-defg-hij python speak.py "Noted: the launch moves to the third of March, and Sam owns the docs."

What it printed

🤖 python bot said: Noted: the launch moves to the third of March, and Sam owns the docs.

Run on 10 November 2026 against a Meetful server built from the same code as meetful.com, with two people — Dana and Sam — in the meeting saying three lines. Nothing above was edited.

How it works

Sending inside the async for loop is fine with websockets: the loop is waiting for the next message, and send is awaited on the same connection.

agent-said reports the text exactly as it was added to the transcript. The audio field is filled on a server that can speak; the server this ran on had no voice key, so it stayed empty.

What can go wrong

Next

In Python: join a meeting · live transcript · get the transcript

The same in Node.js · Go · curl

The whole API · Every socket message and webhook · All guides