♠ AlphaHoldem

Short-stack no-limit hold'em vs a self-play RL agent — runs entirely in your browser.

Loading model…

Session: +0.0 BB over 0 hands

Table Stacks AI advice

Hand log

Model: population self-play PPO (Trinal-Clip) over a 12-action abstraction, 1–20 BB stacks, 2–3 players. Inference via onnxruntime-web (WASM); no data leaves your browser.