pokerai-bet 0.1.0

Official Python client for the Pokerai API — solver-grade GTO strategy for 6-max No-Limit Hold'em. Auto-generated from the OpenAPI spec.

pokerai-bet

Official Python client for the Pokerai API — solver-grade GTO strategy for 6-max No-Limit Hold'em, over HTTP. Typed, httpx-based, auto-generated from the OpenAPI spec so it always tracks the live API.

Get a free API key at https://pokerai.bet/login. Docs: https://pokerai.bet/docs.en

pip install pokerai-bet   # distribution name; import as `pokerai`

Quickstart

Preflop GTO strategy for a hand facing an action line:

from pokerai import AuthenticatedClient
from pokerai.api.lookup import preflop_strategy
from pokerai.models import (
    PreflopRequest,
    PreflopRequestPositions,
    PreflopRequestPreflopActionsItem,
    PreflopRequestPreflopActionsItemAction as Act,
)
from pokerai.models.position import Position

client = AuthenticatedClient(base_url="https://pokerai.bet", token="gto_your_key")

req = PreflopRequest(
    hole_cards="AhKh",
    positions=PreflopRequestPositions(hero=Position.MP),
    preflop_actions=[
        PreflopRequestPreflopActionsItem(position=Position.SB, action=Act.SMALL_BLIND, amount=0.5),
        PreflopRequestPreflopActionsItem(position=Position.BB, action=Act.BIG_BLIND, amount=1),
        PreflopRequestPreflopActionsItem(position=Position.UTG, action=Act.RAISE, amount=3),
    ],
)

result = preflop_strategy.sync(client=client, body=req)
print(result)   # situation="Raise", strategy=[{action:"raise", frequency:1.0, amount_bb:9, ...}]

Layout

Every endpoint is a module under pokerai.api.<tag> with four functions:

function	blocking?	returns
sync	yes	parsed model (or None)
sync_detailed	yes	full Response (status + parsed)
asyncio	no	parsed model
asyncio_detailed	no	full Response

Tags: lookup (preflop + flop presolved), solver (real-time turn/river), range (range updates). Request/response models live in pokerai.models. Auth is your API key via AuthenticatedClient(token=...) (sent as Authorization: Bearer).

Not into constructing model objects?

This client is fully typed but verbose (it's generated). If you'd rather drive the API from an LLM agent, see @pokerai/mcp — the same API as MCP tools. Or call the HTTP endpoints directly; see the reference.

MIT licensed.