From Kalshi Agent to Algorithmic Prediction-Market Trader
Start free with Kalshi Agent — a ~200-line Python Kalshi bot you can clone, run, and read in a single sitting. Take Foundation, then pick your track: Operator (Kalshi bot), Quant (live strategies), or Builder (multi-strategy portfolios and agent orchestration).
The full bot operator journey — enter at any point
Why This Exists
Start with Kalshi Agent
Clone something tiny that works. Read every line. Then graduate through Foundation and the three tracks.
uv run python kalshi_agent.py. Dry-run by default — every trade logged to kalshi_agent.db.kalshi_agent.py, a SQLite logger, and a strategy function. No magic — every line is yours to understand.Foundation
7 modules · Required before any track · Open source
Module 0
How ATC works, the three-track model, and what you ship at each graduation gate.
Module 1
Kalshi's REST and WebSocket APIs, binary contract mechanics, implied-probability math, and pykalshi as your Python client — the foundation every operator and quant builds on.
Module 2
Regulated exchange growth, event-contract liquidity expansion, and why quant edges in mispriced probabilities are still wide open for systematic traders.
Module 3
Claude Code or Cursor as your copilot — project context, permission models, and verification discipline.
Module 4
Python + UV, Git, and the repos you'll work from throughout the program. No blockchain tooling required.
Module 5
Generating API credentials, RSA-PSS request signing, demo vs production environments, and the security posture operators and quants actually use.
Module 6
Drawdown, position limits, bankroll sizing, and why risk gates exist before you trade real capital on live markets.
Three Tracks
Foundation → Track A (Operator) → Track B (Quant) → Track C (Builder). Each track has a graduation gate with a public artifact.
The Stack
Every tool is open source. Every layer is documented.
Kalshi Agent
Minimal open-source Python Kalshi bot — start free, graduate to production
Claude Code / Cursor
AI coding agents — your operator and quant copilots
Kalshi API
REST + WebSocket order-book streaming, RSA-PSS auth — the primary venue
pykalshi
Async Python client with local order-book mirroring and pandas integration
PMXT
CCXT of prediction markets — unified multi-venue routing across Kalshi and Polymarket
Risk controls
Drawdown circuit breakers, Kelly sizing, sector caps
Python 3.12 + UV
Modern Python with Rust-speed tooling
SQLite
Zero-config database for trade logging
VPS / DigitalOcean
24/7 bot deployment with tmux/screen on Ubuntu droplets
Who This Is For
You belong here if:
- • You're a developer comfortable with AI coding tools (Claude Code, Cursor, Copilot)
- • You want to build Kalshi trading bots and deploy algorithmic strategies responsibly
- • You're drawn to prediction markets and quantitative event-contract trading
- • You want to understand the full stack: Python bot → strategy → live markets
- • You value sovereignty: infrastructure you own, public artifacts you ship, no gatekeepers
- • You want to be in the room with quants and operators actively building on Kalshi
This probably isn't your thing if:
- • You're completely new to programming (learn basics first, come back anytime)
- • You're looking for a get-rich-quick scheme or a "magic strategy"
- • You want someone to hand you a profitable strategy without doing the work
- • You're uncomfortable managing API keys, risk, or live trading capital
- • You want a passive, watch-videos-and-forget experience
Open Source Knowledge. Operator Community.
You can absolutely do this alone with the open source materials. But if you want to move faster and be around people deploying Kalshi bots and trading strategies, this is the room.
- • Full curriculum outline and structure
- • Kalshi Agent codebase (GitHub, MIT license)
- • Module summaries and learning objectives
- • SKILL.md, CLAUDE.md, and all documentation
- Everything you need to go solo is free and open source.
- • Full module walkthroughs with commentary and context
- • Weekly live sessions — bot deployment, strategy design, quant ops
- • Community code reviews on Kalshi Agent forks and strategy repos
- • Direct access to Fodé and experienced quants
- • Private channels for real-time discussion
- • Ongoing experiments, strategy performance data, and prediction-market findings
- The curriculum tells you what to learn. The community is where operators build together.
FAQ
What's Kalshi Agent?
Kalshi Agent is a deliberately minimal Kalshi trading bot — ~200 lines of Python. Clone it, run a dry-run trade in minutes, read every line. The same repo grows into the production bot stack taught in Track A — full risk management, WebSocket streaming — when you're ready for live capital.
Do I need coding experience?
Basic programming skills, yes. You should be comfortable with the command line and able to read Python code. But you don't need to be an expert — most of us use AI tools like Claude Code or Cursor to build and iterate. If you can describe what you want in English and verify the output, you'll be fine.
How long does the curriculum take?
Kalshi Agent on its own takes about 30 minutes. Foundation (7 modules) takes 2-3 weeks at 5-10 hours/week. Track A (Operator) is another 4-6 weeks. Track B (Quant) is 8-12 weeks including backtesting and live strategy deployment. Track C launches Q4 2026. It's self-paced.
What's the difference between the Operator track and the Quant track?
Track A (Operator) is infrastructure-first: you ship a production Kalshi bot that runs 24/7 with RSA-PSS auth, REST + WebSocket, and risk controls. Track B (Quant) is strategy-first: you build, backtest, and deploy real strategies — stat arb, NLP sentiment, Safe Compounder EV edge — running live. Most students do Foundation → Track A → Track B.
Do I need blockchain experience?
No — the entire curriculum is Python-first. Track A and Track B use the Kalshi REST + WebSocket API, pykalshi, and standard quant tooling. No blockchain development required.
Is this actually open source?
The curriculum outline, Kalshi Agent codebase, and all documentation are MIT-licensed and on GitHub. Fork it, improve it, teach from it. ATC membership gets you into the community — live sessions, strategy discussions, code reviews, and ongoing experiments.
Will this make me money?
This is a community of builders, not financial advisors. Prediction-market trading involves real risk of real loss. The curriculum teaches technical infrastructure: trading bots, strategy development, risk controls, and operational discipline. Whether any strategy profits depends on market conditions, your risk management, and your judgment.
Who is this for?
Laid-off FAANG engineers looking for a sovereign skill stack. Quants leaving prop shops who want prediction-market infrastructure. Finance professionals moving into event-contract trading. Developers who've traded manually but never built a bot. If you're intellectually curious, technically literate, and want to ship public artifacts — this is your room.
Ready to Build Your First Bot?
Start free with Kalshi Agent. Graduate through Foundation and three tracks. Deploy strategies on Kalshi. Build in public — no permission needed.