Blog · Bac Bo Live

Bac Bo with Python: why your scraper dies, and the fix

Selenium, the casino's websocket, third-party logins: why Bac Bo scrapers break, what they cost to maintain, and how to migrate the same bot to an API.

Published on October 18, 2026 · 4 min read

Almost everyone who builds a Bac Bo bot starts the same way: they search for "bac bo python" or "bac bo github", find a project that opens the game in an automated browser, reads the results off the screen and sends them to Telegram. It works for the first week. Then the outages start, and the bot that was supposed to run by itself becomes a daily job. This article explains why that happens and how to get out of the cycle.

The three most common ways to scrape Bac Bo

1. An automated browser (Selenium or Playwright)

The program opens the bookmaker in a code-controlled Chrome, logs in, enters the table and reads the results off the page. This is what most public projects do.

Where it breaks: the operator changes the layout and the selector stops finding the result, the login expires in the middle of the night, an anti-bot challenge shows up, Chrome eats memory until the server freezes, and the Chromedriver version stops matching Chrome after an update.

2. The casino's websocket

A step further is opening the developer tools, finding the real-time connection that feeds the table and reading the messages directly, with no browser.

Where it breaks: the connection requires a token from a logged-in session, which expires, the message format changes without warning, and many operators block connections that do not come from a real browser or that come from server IPs.

3. Somebody else's statistics site

The most tempting shortcut is to scrape a site that already shows the results, or to use a third-party service's login.

Where it breaks: you depend on something you do not control. When that site goes down, changes or blocks your IP, your bot goes down with it. And not every site records every round: in our comparison, the public list of a well-known site was leaving out about 20% of the rounds.

The cost that does not show up at the start

  • Lost rounds. Every outage opens a gap in the history, and the group notices when the bot skips results.
  • A bigger server. An automated Chrome needs far more memory than a script that only makes HTTP requests, and the hosting plan goes up.
  • Your time. Every change on the site becomes a night of maintenance, and it generally happens when you are busy.
  • Terms of use. Automating access to a bookmaker or a third party's site usually goes against their terms, and the account the bot uses can be blocked.

That cost applies to any scraper, not just in Bac Bo, and it is detailed in scraping or an API.

Migrating the same bot to an API

The good news is that the Telegram part, the messages and your bot's rules all stay the same. Only the function that fetches the result changes. In a typical Selenium bot, it looks something like this:

before.py

from selenium import webdriver
from selenium.webdriver.common.by import By

browser = webdriver.Chrome()
browser.get("https://bookmaker.example/bac-bo")
# ... log in, wait for the table to load, switch iframes ...

def latest_result():
    element = browser.find_element(By.CSS_SELECTOR, ".history .item:last-child")
    return element.get_attribute("class")

With the Bac Bo Live API, the same function becomes:

after.py

import os
import requests

def latest_result():
    r = requests.get(
        "https://api.lynxmetrics.com/v1/bacbo-live/latest",
        headers={"x-api-key": os.environ["LYNX_KEY"]},
        timeout=10,
    )
    r.raise_for_status()
    return r.json()["data"]

No browser, no login and no selectors. The response already carries both dice for each side, the totals, the winner and the tie multiplier, in a format that does not change when somebody's site changes. If you want the complete bot, from BotFather to hosting, the walkthrough is in how to build a Bac Bo Telegram bot.

What if the API goes down?

That is a fair question. On our side, the table reader has automatic alerts when it stops, and every round read is stored in the database. If your bot is offline for a while, it loses nothing: the after parameter on /results returns every round that ended after the last one you have, in order, and the bot catches up by itself when it comes back. With the stream, a short outage is recovered automatically, as explained in Bac Bo results in real time.

Quick comparison

Your own scraperBac Bo Live API
Needs a login at a casinoYesNo
Browser on the serverUsually, yesNo
Breaks when the site changesYesNo, the format is stable
HistoryOnly what your bot sawEverything since 14/10/2026
Recovers lost roundsNoYes, with after
CostA bigger server plus maintenanceFrom R$ 21.90 a month

You can test the switch for free: the 2-hour trial gives you a key with Pro limits, no card, and the code above works with it.

Frequently asked questions

Is there a Bac Bo project on GitHub that works?

There are several, almost all of them using an automated browser or depending on a third-party site's login. They work until the site changes, the session expires or the IP gets blocked, which is why they need constant maintenance.

Can I get Bac Bo results through the websocket?

You can, as long as the session token stays valid and the message format does not change. In practice, the connection requires an account logged into a casino and breaks often, besides possibly going against that casino's terms of use.

How long does it take to swap the scraper for the API?

Usually minutes. Only the function that fetches the result changes, and the rest of the bot, such as messages and rules, stays the same.

For adults 18 and over only. LynxMetrics is not affiliated with Evolution or with the casinos that offer Bac Bo Live, and it only sells data and stats, never predictions or betting tips, because every round is random and past results do not predict the next one.

Read also