Free Polymarket Open Source Tools and Trading Bots on GitHub

Serious Polymarket traders look at data before they place a bet. They pull full trade histories, test ideas against past markets, and watch how winning wallets move. The open source Polymarket tools below help you do that work without paying for a subscription. Every repo here has active code and readable docs. Use them to research faster, size positions with more care, and copy less noise.

• Polymarket Tools

Safety first. Run any repo inside a virtual machine before you trust it. Make a fresh Polymarket wallet and fund it with a small amount for tests. Some public repos hide drainers that steal private keys, or malware that reads your local files. A throwaway wallet and a sandbox keep your main funds out of reach.

Polymarket Analytics and Historical Data Tools

Good trades start with good records. These Polymarket analytics tools give you the raw history and the dashboards to read it.

• Polymarket Analytics Tools

Largest Polymarket Historical Dataset (107 GB)

Five researchers built and maintain the biggest public Polymarket dataset. It covers more than 1.1 billion trades across roughly 268,000 markets, and the full set reaches 107 GB of clean Parquet files. You get order filled events, complete trade records, market metadata, unified yes token views, and user level behavior splits. The pipeline reads straight from Polygon contracts and the Gamma API, and daily updates keep it current. The quant file suits price studies and forecasting. The users file works for wallet PnL tracking and pattern discovery. Hugging Face mirrors host download copies, and an MIT license covers both research and commercial use.

GitHub: https://github.com/SII-WANGZJ/Polymarket_data

Historical Market Backtesting Simulator

This backtester builds on the Nautilus Trader engine and adds Polymarket adapters and data loaders. You replay real historical markets and measure equity curves, drawdowns, Sharpe ratios, and Brier scores. Multi market portfolios run in one pass, and you can tune parameters with tools like Optuna. It models fees, slippage, and maker rebates, so your results match live conditions more closely. Live sandbox hooks already support short interval BTC markets. Pair it with the dataset above to test three or four ideas before you risk a dollar.

GitHub: https://github.com/evan-kolberg/prediction-market-backtesting

Trader Behavior Analyzer (polybot)

Point this bot at any Polymarket wallet and it reads that trader's full history. It scores repeated patterns and shows which strategies the wallet runs. ClickHouse storage and Grafana dashboards make the output easy to scan. Replication scores tell you which wallets are worth copying.

GitHub: https://github.com/ent0n29/polybot

Cross Platform Market Search (PMXT)

PMXT works like a single interface across prediction market platforms. Think of it as the CCXT of prediction markets. One set of calls reaches Polymarket, Kalshi, and several others. Python and TypeScript SDKs ship ready to use. You can run it hosted or self hosted. AI agents connect through the MCP protocol with no extra glue code.

GitHub: https://github.com/pmxt-dev/pmxt

Wallet History Collector (collectmarkets2)

Give this tool an address and it pulls the complete activity list, drops duplicates, groups trades by market, and exports clean CSV files. A cyclic monitor keeps the data fresh. Charts and summary stats show up in a simple terminal view.

GitHub: https://github.com/txbabaxyz/collectmarkets2

Open Source Polymarket Trading Bots

Research points you at an edge. These Polymarket trading bots act on it. Most support paper trading, so you can measure results before you spend real money.

• Polymarket Trading Bots
• Polymarket Trading Terminals

CloddsBot (118+ Strategies)

CloddsBot covers the most ground. A Cambridge computer science student built it during a Solana hackathon and kept adding to it. The agent runs on Polymarket, Kalshi, Binance, Hyperliquid, and several DEXes. More than 118 strategies sit ready, including momentum, mean reversion, penny clipping, expiry fade, DCA, smart routing, and whale tracking. Polymarket logic handles five minute, fifteen minute, one hour, and daily rounds. The risk engine adds Kelly sizing, daily loss limits, and volatility regime detection. It runs self hosted and keeps working while you sleep.

GitHub: https://github.com/alsk1992/CloddsBot

Liquidity Reward Manager (LP Tool)

This tool never opens new orders. It watches your existing limit orders and nudges them to stay inside the reward bands. Coarse and fine tick rules keep you in range for the liquidity incentives. Telegram commands and a small web dashboard let you change rules without editing code. Rust and Python versions both exist.

GitHub: https://github.com/lihanyu81/polymarket_lp_tool

Smart Money Copy Trading Bot

This bot scans leaderboard wallets and filters hard on win rate, profit factor, and steady performance. It then builds an automatic copy list. Arbitrage, dip catching, and direct order tools sit alongside the copy engine. Risk stops run on daily, monthly, drawdown, and total limits. Dry run mode lets you check its behavior first.

GitHub: https://github.com/MrFadiAi/Polymarket-bot

Rust Trading Toolkit

This toolkit packs ten production strategies onto one execution core written in Rust. Copy trading, cross market arbitrage, spread farming, sports execution, resolution sniping, order book imbalance, and market making all share the same low latency engine. Memory stays under 50 MB. Circuit breakers and depth guards stop the bot from overreaching. Adapters already cover Polymarket, Kalshi, and more venues.

GitHub: https://github.com/HarrierOnChain/Prediction-Markets-Trading-Bot-Toolkits

Polymarket Kalshi Arbitrage Bot

This bot hunts price gaps between Polymarket and Kalshi. The idea is plain. When yes on one side plus no on the other costs under one dollar after fees, the bot locks the spread. Fast order book feeds and quick execution matter more than clever math here.

GitHub: https://github.com/Composio-HQ/polymarket-kalshi-arbitrage-bot

Weather Trading Tools for Temperature Markets

Temperature contracts stay among the cleaner edges on Polymarket and Kalshi. Public forecasts carry known biases. Bots that correct those biases keep finding mispriced buckets.

Machine Learning Weather Model (Kalshi)

This system trains an XGBoost model on several years of history from multiple weather providers. It learns how each source over or under predicts in specific cities, then adjusts the next forecast before it compares that number to market prices.

GitHub: https://github.com/AruneshDev/Automated-Trading-System-Kalshi-Weather-Model

Real Time Weather Report Bot (PolyWeather)

A Chinese developer built PolyWeather to read METAR and SPECI aviation observations alongside standard forecasts. It produces detailed city and day reports that feed straight into your trading calls.

GitHub: https://github.com/yangyuan-zhen/PolyWeather

Autonomous Weather Bot (Hermes Agent)

The Hermes version runs on its own. It pulls temperature data from several sources and uses a Gaussian bucket method. If the forecast reads 72F, it expects the final reading between 71F and 73F. It compares that range to live Polymarket prices, sizes with fractional Kelly, and sends a signal to Telegram. After each trade it updates its own parameters, so win rate and PnL shape the next cycle.

GitHub: https://github.com/nicolastinkl/hermes_weatherbot

GFS Ensemble Weather Bot

This bot pulls all 31 members of the Global Forecast System for a target city and day. The spread of those runs becomes a probability distribution, which the bot compares against live market prices. A dashboard tracks every signal, fill, forecast, and running PnL.

GitHub: https://github.com/suislanchez/polymarket-kalshi-weather-bot

NWS Kelly Weather Bot

This bot reads the latest official National Weather Service numbers for a chosen city and compares them to current Polymarket prices. It applies Kelly sizing based on edge strength and can alert you or trade on its own. Simulation logs let you review each decision later.

GitHub: https://github.com/MoonsatProtocol/Polymarket-Weather-Bot

AI Powered Tools and Agent Frameworks

These layers sit on top of the data and execution tools. They cut the hours you used to spend reading by hand.

• Ai Polymarket Tools

Multi Agent Trading Dashboard (TradingAgents)

Several specialized models examine the same market at once. One agent checks news flow. Another reads price action and technical signals. A third flags risk. Their group discussion gives you a clearer read than any single model.

GitHub: https://github.com/TauricResearch/TradingAgents

Polymarket MCP Server

This MCP server gives Claude direct access to Polymarket. Forty five tools cover market discovery, analysis, order placement, portfolio tracking, and live monitoring. Safety limits and confirmation steps stay configurable, and a local dashboard shows the state in real time. You can connect your own account and let it place trades.

GitHub: https://github.com/caiovicentino/polymarket-mcp-server

30 Day News Research Skill

This skill pulls the last thirty days of coverage across several platforms and writes a grounded brief. It weights real money odds from Polymarket into the summary, which helps before you take a position.

GitHub: https://github.com/mvanhorn/last30days-skill

Pydantic AI

Pydantic AI gives you the parts to build your own trading agents. Type safe tool calling, structured outputs, and human approval steps make a production agent more dependable.

GitHub: https://github.com/pydantic/pydantic-ai

GPT Researcher

Point GPT Researcher at any market question and it runs a deep, multi source investigation. It returns a full report you can read before you decide.

GitHub: https://github.com/assafelovic/gpt-researcher

Quick Comparison of Polymarket Open Source Tools

How to Use These Polymarket Tools Without Losing Your Shirt

Start with the 107 GB dataset and the backtester. Measure three or four plain strategies against real history. Move to paper trading with one bot only after that. Keep position sizes small until your live curve matches the backtest.

Weather markets reward clean data pipelines more than fancy models. Ensemble forecasts plus basic Kelly sizing already beat most manual guessing. Traders tend to report steadier results on high volume cities when the model edge clears about eight to ten percent and volume stays healthy.

AI agents work best on top of solid infrastructure. Feed them clean data and firm risk limits. Do not hand them full wallet access on day one.

The open source scene around Polymarket moves fast, with new commits most weeks. Star the repos you depend on so updates reach you, and fork anything you plan to change a lot.

Final Word and Next Step

Traders who treat these tools as research instruments keep their capital far longer than the ones chasing a money printer. Measure everything and keep sizes small.

Ready to start? Bookmark this page, clone the two starter repos (the dataset and the backtester), and run your first backtest this week inside a sandbox. Once a strategy proves itself on real history, move it to paper trading with one bot before you risk live funds.

• Polymarket Tools And Bots

Disclaimer: This post shares information about open source Polymarket tools and is not financial advice. Prediction market trading carries a real risk of loss. Check that Polymarket, Kalshi, and this kind of activity are legal where you live before you trade.