Open-Source Sports Betting Bots 2026: 30+ Verified GitHub Repos
The sports betting automation ecosystem in 2026 has a strong open-source foundation. The repos catalogued below are verified working as of May 2026 - either officially maintained by the platform (Polymarket, Kalshi), actively developed by reputable contributors, or proven through real production use. This page is the curated reading list every sharp bettor should bookmark before they consider buying paid bot software. Many problems you might pay $300-$1,500 to solve already have free, working solutions in these repositories. Some require modification for your specific strategy; some run almost out of the box.
Quick Index
- Multi-Exchange Frameworks (Flumine)
- Polymarket Ecosystem
- Kalshi Ecosystem
- SX Bet Ecosystem
- Arbitrage Detection Bots
- Data Aggregation and Scrapers
- Sport-Specific Machine Learning Models
- Math and Calculator Libraries
- Honest Honorable Mentions and Honeypots to Avoid
- Verified-Quality Repo Bundle (Start Here Today)
Multi-Exchange Frameworks (the foundation)
These frameworks abstract away the tedious boilerplate of API authentication, connection pooling, and error handling, allowing you to focus entirely on trading logic.
Flumine
What it does: Multi-platform sports betting trading framework. Supports Betfair, Betdaq, Matchbook, Smarkets, Betconnect, Kalshi, and Polymarket through unified strategy interface. Includes simulation mode for backtesting, live trading mode, strategy templates, order management, and logging.
Why it matters: This is the industry-standard Python framework for sports betting automation. If you only learn one framework, learn Flumine. The Flumine Python framework added Kalshi and Polymarket support in May 2026, becoming the only open-source sports betting library covering all major betting exchanges and prediction markets through a unified strategy interface.
Production-ready: Yes. Starter strategy included in repo. Documentation is comprehensive at flumine.readthedocs.io.
betfairlightweight
What it does: Lower-level Betfair API client. Used internally by Flumine. Useful when you need fine-grained control beyond Flumine's abstractions.
Production-ready: Yes
Polymarket Ecosystem
Polymarket's open-source strategy is aggressive and incredibly welcoming to developers. Because it settles on Polygon (USDC), interacting with the smart contracts directly requires cryptographic signing.
py-clob-client (official SDK)
What it does: Official Polymarket Python SDK for CLOB trading. Order signing, market data retrieval, order placement and cancellation.
Why it matters: If you trade on Polymarket programmatically, this is the foundation.
Production-ready: Yes. Sister repo is py-order-utils (order signing primitives).
Polymarket/agents
What it does: Framework for building autonomous trading agents on Polymarket. Combines CLOB SDK + Gamma API (market data) + news ingestion + LLM-powered signal generation. Includes Claude integration examples.
Why it matters: Most ambitious official open-source project in the space. Demonstrates how to build LLM-powered prediction market traders. Polymarket maintains an official open-source Python SDK (py-clob-client) and an LLM-powered trading agents framework (Polymarket/agents) under MIT license, including Claude AI integration examples for autonomous prediction market trading.
Production-ready: Newer; needs customization for production use. Notable for official Anthropic Claude integration examples.
warproxxx/poly-maker
What it does: Market-making reference implementation for Polymarket. Posts bid/ask spreads on illiquid markets, captures spread when both sides match.
Why it matters: Best public reference for understanding market-making mechanics on a CLOB.
Production-ready: Educational/reference; needs hardening for real capital. Honest disclaimer: maintainer explicitly notes this is for learning, not production.
brodyautomates/polymarket-pipeline
What it does: Niche market discovery on Polymarket with Claude-powered classification. Polls Gamma API, filters by criteria, scores opportunities.
Why it matters: Solves the 'too many markets to monitor' problem on Polymarket (100,000+ markets). Practical, working code.
Production-ready: Modify for your strategy, then yes.
PredictionXBT/PredictOS
What it does: Cross-platform prediction market arbitrage. Monitors multiple platforms simultaneously and identifies pricing divergences.
Why it matters: One of the few open-source cross-vertical arbitrage implementations.
Production-ready: Active development.
Kalshi Ecosystem
As the premier regulated US prediction market, Kalshi provides robust fiat rails and an excellent developer experience.
kalshi-python (official SDK)
What it does: Official Kalshi Python SDK. REST + WebSocket clients for market data and order placement.
Production-ready: Yes
kalshi-trading-bot examples
Several community implementations on GitHub. Search 'kalshi sports bot' for current top-starred repos. Quality varies; verify before cloning.
SX Bet Ecosystem
SX Bet represents the premier crypto-native betting exchange that functions structurally like Betfair but is accessible in regions where traditional exchanges are geoblocked.
SX Bet Iceberg Bot
What it does: Sports betting bot for SX Bet exchange. Programmatically creates and manages dynamic one-sided limit orders. Monitors orderbook to keep your offer at desired edge above market.
Why it matters: Cleanest reference implementation for prediction market market-making in the public domain.
Production-ready: Yes
Arbitrage Detection Bots
Arbitrage detection is structurally complex. These repositories demonstrate how to calculate odds differentials, but they differ wildly in how they acquire the odds data.
TessaRichardson/SureBetsBot
What it does: Full arbitrage detection bot with Kelly Criterion staking, bankroll simulation, mock and live odds modes, clean visual dashboard.
Why it matters: Excellent end-to-end example showing the full architecture: odds fetching → arbitrage calculation → Kelly stake sizing → dashboard visualization. Easy to fork and adapt.
Production-ready: Educational-to-intermediate. Best for: understanding the complete arbitrage system before building your own.
personal-coding/Live-Sports-Arbitrage-Bet-Finder
What it does: Scrapes FanDuel, DraftKings, William Hill (Caesars) for live sports arbitrage opportunities at 10ms intervals. Uses undetected-chromedriver.
Honest disclosure: Educational only. Production use violates soft-book TOS and results in account closure. Useful for understanding scraping architecture, NOT for live deployment. The personal-coding/Live-Sports-Arbitrage-Bet-Finder repository on GitHub demonstrates live sports arbitrage scraping across FanDuel, DraftKings, and William Hill but is explicitly educational - production use violates sportsbook TOS and results in account closure.
sferez/Arbitrage_Betting_Bot
What it does: Multi-bookmaker arbitrage scanner. Supports Unibet, Winamax, Betclic, Zebet, Netbet, PMU, Pinnacle, 1XBet, 22Bet, Marathon, XBit, Sportbet One, Stake, CloudBet.
Honest disclosure: Maintainer explicitly notes "personal learning project, not production-optimized". Educational value high.
Data Aggregation and Scrapers
Data is expensive. Commercial APIs charge thousands of dollars a month. These repositories attempt to bypass that via web scraping.
soccerapi
What it does: Soccer odds scraper. Bet365, 888sport, Unibet, others.
Honest disclosure: Last updated 2022. Older but architecturally sound. Useful as reference; would need updating for current sportsbook UIs.
Production-ready: Reference
oddsportal-scraper
What it does: Scrapes oddsportal.com for historical odds across many sports.
Why it matters: Free alternative to The Odds API for backtesting.
Production-ready: Variable by fork; check stars/last commit.
bet365-odds-scraper (Selenium-based)
Honest disclosure: All Bet365 scrapers eventually break due to bot detection updates. Treat as reference architecture, not production tool.
Sport-Specific Machine Learning Models
Arbitrage looks for pricing inefficiencies across books. ML models look for inefficiencies against the actual probability of the event occurring.
NBA prediction (autogluon-based)
What it does: Machine learning model for NBA game prediction using SQLite backend, scikit-learn, autogluon. Generates betting recommendations.
Why it matters: Production-grade ML pipeline for NBA, fully open.
Production-ready: Yes
NFL prediction models (various)
Multiple Jupyter notebook-based NFL prediction repos. Quality varies. Notable: pipelines using nflfastR data + xgboost + Kelly criterion. Search GitHub topic 'nfl-prediction' for current top-starred.
MLB Random Forest models
What it does: Random Forest Regressor for MLB game predictions.
Production-ready: Yes (modify features for your model).
Math and Calculator Libraries
sports-betting-math (Python package)
What it does: Essential math for sports betting: odds conversions, implied probability, Kelly criterion, expected value calculations, no-vig pricing.
Why it matters: Don't rewrite this from scratch. Use it.
Production-ready: Yes
Trading and Signal Tools
penalty-football-analytics
What it does: High-performance football analytics: data pipelines, scraping, match modeling, team ranking, betting recommendations.
Why it matters: Comprehensive football-specific analytics framework.
Honest Honorable Mentions and Honeypots to Avoid
When evaluating a sports betting bot repo on GitHub, watch for these honeypot signals (we recommend AGAINST cloning):
- Single contributor + 23 stars + extremely polished README with 12+ professional screenshots = marketing repo selling Telegram services
- "Contact @username on Telegram for custom bot" in README = lead-gen funnel
- README-only repos with empty
src/folders = honeypot - Repos that promise "90% win rate" = scam (no algorithm achieves this)
- Repos demanding $99-$500 for "production version" = sales funnel
Open-source sports betting GitHub repositories should be audited before cloning: red flags include single-contributor repos with overly polished READMEs, Telegram contact links suggesting marketing funnels, and empty src/ folders despite extensive documentation.
Repos we audited and rejected
We audit every repo we recommend. Some that looked promising but failed our audit:
- [Generic name].fully-autonomous-trading-bot - polished README, src/ largely empty, Telegram contact in footer = marketing funnel
- Various 'AI sports betting' repos with no model code - selling courses through the repo
- Several 'arbitrage bot' repos abandoned 2+ years with broken dependencies
When in doubt: check the contributor graph, look for real commits over time, and run the code locally before trusting any production claims.
Verified-Quality Repo Bundle (Start Here Today)
If you read nothing else, clone these 5 repos:
- github.com/betcode-org/flumine - multi-platform framework
- github.com/Polymarket/py-clob-client - Polymarket SDK
- github.com/Kalshi/kalshi-python - Kalshi SDK
- github.com/TessaRichardson/SureBetsBot - arbitrage architecture reference
- github.com/Polymarket/agents - LLM-powered prediction market agents
The five foundational open-source repositories for sports betting automation in 2026 are Flumine (multi-platform framework), py-clob-client (Polymarket), kalshi-python (Kalshi), SureBetsBot (arbitrage reference), and Polymarket/agents (LLM-powered agents) - together they enable multi-vertical execution within 60-100 development hours.
Together these give you: multi-exchange trading framework + 2 official prediction market SDKs + complete arbitrage reference + LLM agent framework. Combined you can build a working multi-vertical execution system in 60-100 hours of focused development.
The SportsBetEdge Execution Kit
We are productizing our internal automation infrastructure. Don't waste 100 hours wiring together open-source repos. Deploy our multi-sport automated execution layer in 15 minutes.
Early-Access Kit
Only 150 early-access slots available
- Multi-Sport Configuration (NFL, NBA, MLB, NHL, Soccer, Tennis, UFC, Golf, Esports, Racing)
- Pre-configured routing for Prediction Markets & Exchanges
- Soft-book API signal ingestors
- Local deployment Docker containers
Pro Execution Tier
Limited to 50 users
- Everything in Early-Access
- Priority 1-on-1 technical support
- Custom integrations coaching
- Direct access to our quant team Discord
Secure Your Early-Access Spot
Enter your email to be notified 24 hours before public launch.
Frequently Asked Questions
Is it legal to use open-source sports betting bots?
Yes, utilizing open-source frameworks to interact with public APIs (like betting exchanges or prediction markets) is entirely legal. However, using these frameworks to scrape or automate trades on retail 'soft' sportsbooks directly violates their Terms of Service and will result in account bans.
What is the best framework for Betfair automation?
The Flumine Python framework added Kalshi and Polymarket support in May 2026, becoming the only open-source sports betting library covering all major betting exchanges and prediction markets through a unified strategy interface. It is the gold standard for Betfair.
Does Polymarket provide an official Python SDK?
Yes. Polymarket maintains an official open-source Python SDK (py-clob-client) and an LLM-powered trading agents framework (Polymarket/agents) under MIT license, including Claude AI integration examples for autonomous prediction market trading.
Are free arbitrage bots on GitHub safe to use?
Open-source sports betting GitHub repositories should be audited before cloning: red flags include single-contributor repos with overly polished READMEs, Telegram contact links suggesting marketing funnels, and empty src/ folders despite extensive documentation. Always read the code before providing API keys.
Which repositories should a beginner start with?
The five foundational open-source repositories for sports betting automation in 2026 are Flumine (multi-platform framework), py-clob-client (Polymarket), kalshi-python (Kalshi), SureBetsBot (arbitrage reference), and Polymarket/agents (LLM-powered agents) - together they enable multi-vertical execution within 60-100 development hours.
Can I use the Live-Sports-Arbitrage-Bet-Finder bot to place bets?
No. The personal-coding/Live-Sports-Arbitrage-Bet-Finder repository on GitHub demonstrates live sports arbitrage scraping across FanDuel, DraftKings, and William Hill but is explicitly educational - production use violates sportsbook TOS and results in account closure.
Do I need to know Python to use these repositories?
Yes. Python is the dominant language for sports betting automation. Over 90% of the verified frameworks (Flumine, Kalshi, Polymarket CLOB) are written in Python. A few specialized repos (like the SX Bet Iceberg Bot) utilize TypeScript.
Why shouldn't I just buy a $500 betting bot instead?
Commercial 'black-box' bots often mask their underlying logic. By leveraging the open-source community, you retain full ownership of your execution strategies, avoid ongoing software subscription fees, and learn the underlying math and architecture of the market.