How to follow a channel in Python with the Wapito WhatsApp API

Before you start

• pip install requests
• os.environ["WAPITO_TOKEN"]
• A channel with a WhatsApp number linked to it, and its API token. Create one in the dashboard — the authentication guide shows where the token goes.

The whole script

Every step above in one runnable file. Save it as follow-channels.py, put your token in the environment, and run it.

"""Follow Channels with the Wapito WhatsApp API.

Search or resolve a link, verify it is the right channel, page back through the history, then handle new posts from the webhook.

Run it with:
    export WAPITO_TOKEN="wpt_..."
    python follow-channels.py
"""
import os

import requests

BASE_URL = "https://api.wapito.com/v1"
TOKEN = os.environ["WAPITO_TOKEN"]

# --- Search the directory ---
url = BASE_URL + "/newsletters/find"

querystring = {"q":"release notes","country":"GB","count":"50"}

headers = {"Authorization": "Bearer " + TOKEN}

response = requests.get(url, headers=headers, params=querystring)

print(response.json())

# --- Resolve an invite link ---
url = BASE_URL + "/newsletters/invite/0029VaAbCdEfGhIjKlMnOp"

headers = {"Authorization": "Bearer " + TOKEN}

response = requests.get(url, headers=headers)

print(response.json())

# --- Read what has been published ---
url = BASE_URL + "/newsletters/120363099887766554@newsletter/messages"

querystring = {"count":"50","offset":"0"}

headers = {"Authorization": "Bearer " + TOKEN}

response = requests.get(url, headers=headers, params=querystring)

print(response.json())

Receive the webhook

FastAPI receiver for messages — it verifies the signature, answers immediately and does the work after. Install it with pip install fastapi uvicorn and save it as webhook.py.

import hashlib
import hmac
import os
import time

from fastapi import FastAPI, Header, HTTPException, Request

app = FastAPI()
SECRET = os.environ["WAPITO_WEBHOOK_SECRET"].encode()
EVENTS = ["messages"]


def verify(raw: bytes, header: str | None) -> bool:
    """Checks the X-Wapito-Signature header: t=<ms>,v1=<hex hmac-sha256>."""
    if not header:
        return False
    parts = dict(p.split("=", 1) for p in header.split(",") if "=" in p)
    ts, sig = parts.get("t"), parts.get("v1")
    if not ts or not sig:
        return False
    if abs(time.time() * 1000 - int(ts)) > 300_000:  # 5 minute clock skew
        return False
    expected = hmac.new(SECRET, ts.encode() + b"." + raw, hashlib.sha256).hexdigest()
    return hmac.compare_digest(expected, sig)


@app.post("/wapito")
async def wapito(request: Request, signature: str | None = Header(default=None, alias="X-Wapito-Signature")):
    raw = await request.body()
    if not verify(raw, signature):
        raise HTTPException(status_code=401, detail="bad signature")
    payload = await request.json()
    if payload["event"] in EVENTS:
        print(payload["event"], payload["data"])
    return {"ok": True}  # answer 2xx fast; do the real work in a queue

Gotchas

• requests does not raise on a 4xx by itself. Call response.raise_for_status() or check response.ok, otherwise a 429 send_rate_limited silently looks like a successful send and your loop keeps hammering the queue.
• json= and data= are not interchangeable. Pass json={...} so requests sets Content-Type: application/json and encodes UTF-8 for you; data= with a dict sends form encoding and the API answers 400 invalid_request.
• The default requests timeout is None, which means a stalled connection hangs your worker forever. Always pass timeout=(5, 30) and treat a read timeout as "maybe delivered" - re-check with the message id before you send again.
• A group id such as 120363000000000000@g.us is a string, not a number. Building it with an f-string from an int drops the precision and you get 404 not_found.
• Emoji and accented text need no extra work in Python 3, but reading a CSV of numbers with the default encoding on Windows will mangle them - open files with encoding="utf-8".

Where to run it

• Google Cloud Run (container, scale to zero, one revision per deploy)
• Render background worker for the sender, Render web service for the webhook receiver
• A 1 GB VPS with systemd for the worker and Caddy in front of the FastAPI receiver
• AWS Lambda + API Gateway for the webhook receiver only (the sender needs a long-lived process for queue spacing)

Pitfalls in Python

• A set of processed ids kept only in memory is lost on restart, and the bridge reposts the whole archive; persist it, or store the newest id and compare ordering, and load it once at startup rather than on every event.
• requests.get with the term concatenated into the URL string breaks on spaces and non-ASCII names; pass params= and let requests encode it, which also keeps a plus sign in a search term from being read as a space.
• json.dump of the id set on every event rewrites the whole file and a crash mid-write leaves it truncated, so the next start finds invalid JSON and starts from nothing; write to a temporary file and os.replace it, or keep the ids in SQLite with a unique constraint.

Frequently asked questions

Related