Siva-Chidambaram12

Siva-Chidambaram12 / kalshi-trading-bot

Public

🏗 AI trading system for Kalshi prediction markets. kalshi trading bot kalshi trading bot kalshi botFeatures Grok-4 integration, multi-agent decision making, portfolio optimization, and real-time market analysis. Educational/research purposes only kalshi trading bot kalshi bot

bot kalshi kalshi-bot kalshi-trading prediction-market
66
508
75% credibility
Found May 17, 2026 at 66 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

This is an AI-powered trading bot for Kalshi prediction markets that uses five different AI models working together as a team. Each model specializes in a different role—one analyzes news, another argues for buying, another argues for selling, a risk manager has veto power, and a final trader makes the call. The system only trades when the models agree, sizes positions using a mathematical formula called Kelly Criterion, and defaults to paper trading so you can test it safely first. It includes hard limits on daily spending, position sizes, and daily losses to prevent catastrophic outcomes. The bot monitors markets in real-time, aggregates news from major sources, tracks all trades in a database, and provides a dashboard to review performance. It's designed for people who want disciplined, emotion-free trading on prediction markets.

How It Works

1
🔍 You discover an AI trading system

You hear about a bot that uses five different AI minds to trade prediction markets while you sleep.

2
⚙️ You connect your accounts

You link your Kalshi trading account and an AI service that powers five different thinking models.

3
🤖 Your AI team goes to work

Five specialized AI agents analyze markets together—a forecaster, news analyst, bull researcher, bear researcher, and risk manager all debate before any decision is made.

4
📊 You watch everything in the dashboard

A colorful dashboard shows every trade, every model opinion, and your paper trading results so you can learn how the system thinks.

5
You choose how to trade
📝
Paper trading mode

Practice with simulated money until you're confident the system works for you.

💵
Live trading mode

Trade with real money, but the bot automatically limits how much it can spend each day.

6
🛡️ Safety features protect you

The bot refuses to trade if models disagree too much, limits each position to 3% of your balance, and stops if you lose more than 10% in a day.

💤 You sleep while it trades

The system monitors your positions, automatically sells at your stop-loss or profit targets, and logs everything so you can review it anytime.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 66 to 66 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is kalshi-trading-bot?

A TypeScript trading system for Kalshi prediction markets that runs five AI models simultaneously to debate and decide on trades. Instead of relying on a single LLM, it orchestrates a forecaster, news analyst, bull researcher, bear researcher, and risk manager that argue it out before any order touches the wire. Position sizing uses the Kelly Criterion, and every decision is logged to SQLite for later analysis.

Why is it gaining traction?

The multi-agent debate architecture is the hook. Most trading bots are a single model prompted to gamble; this one forces disagreement between models and penalizes confidence when analysts diverge. It ships with paper trading by default, a hard daily AI spend cap that physically prevents the router from calling models when the budget is out, and a full CLI for running iterations, checking scores, and reviewing trade history. The Safe Compounder strategy specifically targets high-probability NO positions with a curated skip-list for sports and entertainment markets.

Who should use this?

Prediction market traders who want a disciplined, multi-model framework without building the orchestration layer from scratch. Researchers backtesting ensemble calibration strategies. Developers evaluating whether Kalshi API integration is worth the effort. Not suitable for production deployment given the 66-star count, limited test coverage, and educational disclaimer.

Verdict

At 66 stars with a 0.75% credibility score, this is an early-stage experimental project with more ambition than polish. The architecture is sound and the safety layers are thoughtful, but the codebase lacks the maturity most traders would need to trust real capital. Start in paper mode, review the SQLite logs carefully, and treat it as a research framework rather than a production-ready bot.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.