What is a GTO poker solver API for 6-max NLHE?
A GTO poker solver API lets software request documented strategy data for a specified poker spot. Pokerai API is a hosted HTTP API, not a desktop solver download: it separates presolved strategy lookups from custom-solve requests so training products, review tools, and agents can use JSON responses without embedding local solver software.
Quick facts
| Product | Pokerai API, hosted at pokerai.bet |
|---|---|
| Game scope | 6-max No-Limit Texas Hold'em: presolved preflop and flop workflows plus custom flop, turn, and river solves |
| Core output | Mixed-frequency actions, ranges, nodes, or EVs; benchmark/gameplay results measure an evaluation or simulation instead |
| Data model | Stateless JSON requests and responses with bearer-token authentication |
| Developer surfaces | OpenAPI, JavaScript SDK, Python SDK, MCP server, docs, and live API reference |
| Free tier | 1,000 presolved lookups + 25 real-time solves per month; no credit card required |
Choose the API type by the question you need to answer
Use the strategy API when a product needs a documented strategy result for a spot. Presolved preflop and flop endpoints return mixed frequencies, ranges, tree nodes, or EVs; a hand can fold, call, and raise at different frequencies rather than have one forced action.
Use the custom-solve API only when the product needs to submit a documented flop, turn, or river configuration. It returns a solve handle first; the strategy result becomes readable only after its tree is queryable. Use benchmark or gameplay APIs for agent evaluation or game simulation. Their scores, outcomes, or trajectories are not a GTO recommendation and should not be presented as one.
Minimal strategy call and result semantics
For one presolved preflop strategy lookup, send a documented spot to POST /v1/gto/preflop:
curl -s https://pokerai.bet/v1/gto/preflop -H "Authorization: Bearer $POKERAI_API_KEY" -H "Content-Type: application/json" -d '{"hole_cards":"AsKs","positions":{"hero":"BTN"},"preflop_actions":[{"position":"SB","action":"small blind","amount":0.5},{"position":"BB","action":"big blind","amount":1},{"position":"UTG","action":"fold"},{"position":"MP","action":"fold"},{"position":"CO","action":"fold"}]}'Read the documented response as a distribution: action frequencies describe the strategy mix for that request, while quota fields report usage rather than strategy quality. For a custom later-street spot, POST /v1/gto/solver schedules the work; follow the solver lifecycle to poll the tree and read a returned node token.
Coverage, quotas, SDKs, and MCP
The public GTO surface covers presolved preflop and flop requests, custom flop/turn/river solve scheduling and reads, range transforms, projected ranges, and EV queries. Check the API Reference or OpenAPI snapshot for the current schemas rather than assuming an undocumented game format, field, or latency.
The Free tier includes 1,000 presolved lookups + 25 real-time solves per month. JavaScript/TypeScript and Python SDKs expose the same public contract for application code; the MCP server exposes it as tools for compatible agents. An agent still needs validated inputs and human review for consequential explanations.
When not to use this page
Choose the hosted API vs desktop solver comparison when the decision is about research control, local tree construction, and integration tradeoffs. Pokerai does not provide a downloadable desktop solver; it provides documented strategy data through hosted HTTP endpoints. This guide is the primary answer for what a poker solver API does; it is not a desktop-solver ranking or a performance claim.
Pokerai API is for training, coaching, hand review, study, and research. Real-time assistance at real-money tables is prohibited. Do not turn a strategy, custom-solve, benchmark, or gameplay response into live-table action advice.
Endpoint map
Use this map when deciding which API surface powers a feature.
| Endpoint | Product task | Use it for |
|---|---|---|
POST /v1/gto/preflop | Grade one preflop hand | Return the GTO action frequencies for a specific hole-card hand and action line. |
POST /v1/gto/preflop/range | Render a 13x13 range grid | Return fold/call/raise frequencies for every hand class in one spot. |
POST /v1/gto/flop/tree | Build a flop decision tree | Fetch presolved flop ranges, actions, and node tokens for a board and formation. |
POST /v1/gto/flop/node | Read one flop decision | Fetch the strategy at a selected node without reloading the whole tree. |
POST /v1/gto/solver | Solve a custom later-street spot | Submit a flop, turn, or river configuration to the real-time solver pool. |
POST /v1/gto/solver/tree | Poll a custom solve | Check readiness and select a returned node token after a submitted solve. |
POST /v1/gto/solver/node | Read one solved node | Fetch strategy for one node in a solved custom spot. |
POST /v1/gto/evs | Explain action quality | Return per-hand expected values for actions at a solved node. |
Related resources
- Developer docs — authentication, quickstart, SDKs, quotas, errors, and endpoint behavior
- Interactive API reference — the live OpenAPI reference for every endpoint
- OpenAPI specification — the machine-readable English API contract
- Build a GTO trainer — a runnable tutorial that makes a real preflop strategy call
- Custom-solve lifecycle — schedule, poll, and read one submitted solver request
- Hosted API vs desktop solver — a transparent integration and research choice
- Counterplay case study — a real iOS trainer built on the API
- llms.txt — the concise LLM entry point for Pokerai API