AI That Masters Poker

Agent2Poker is an open research project training reinforcement learning agents to play Texas Hold'em at superhuman levels using real-time decision modeling and opponent profiling.

Read the Paper View Demo

94.7%

Win rate vs. amateur pool

12M+

Hands trained on

3.2ms

Average decision latency

v0.9.1

Latest model release

🧠

Deep RL Engine

Built on PPO with custom reward shaping for long-horizon poker strategy, including bluff detection and pot-odds calculation.

📡

Real-time API

Stream live game states to our inference endpoint and receive action probabilities within milliseconds via gRPC streaming.

🔍

Opponent Modeling

Adaptive profiling system tracks betting patterns across sessions to dynamically adjust strategy against each player type.

📊

Hand History Analytics

Upload hand histories to get EV analysis, leak detection, and personalized recommendations powered by our trained agent.

🔐

Privacy First

All inference runs on-device or through encrypted transport. No hand data is stored on our servers beyond the active session.

⚙️

Open Weights

Model weights for our 6-max NLHE agent are publicly available under MIT license. Fine-tune for your specific game format.

Live Agent Output

[agent] Hand #48291 — 6-max NL Hold'em /
Hero: A♠ K♦ | Position: BTN | Stack: 12.50
Board: 9♣ K♥ 2♦ | Pot: 8.00 | Action: Hero
[model] EV analysis: Bet 67% pot (+.42 EV) | Check (-bash.18 EV)
[action] BET 2.00 — confidence: 0.91

Villain calls. Turn: 7♠ | Pot: 2.00
[model] Range advantage: 68% | Equity: 74.3%
[action] BET 8.00 — confidence: 0.87

Villain folds.
[result] +2.00 | Session EV: +.71/100 hands