Pricing / monthly plans

Start free. Pay when real-time solves matter.

Presolved lookups and real-time solves are metered separately. Every plan gives you both, so you can build against MelaSolver GPU before committing to production volume. Persistent postflop cache is available only on active API Builder and Pro plans. Other API features keep the same integration contract.

Plans

Choose your capacity and cache access.

Persistent postflop cache is available only on active API Builder and Pro plans. Other API features keep the same integration contract.

Free

$0

Make your first calls and feel both response paths.

  • 10 presolved lookups
  • 5 real-time solves
  • no card required
  • max 12G decision tree
    ?If a tree exceeds this limit, intelligent pruning may be applied or the solve may fail.
Get API key

Starter

$3/mo

Build a full trainer and run every tutorial.

  • 1,000 presolved lookups
  • 25 real-time solves
  • for side projects
  • max 12G decision tree
    ?If a tree exceeds this limit, intelligent pruning may be applied or the solve may fail.
Choose plan

Builder

Recommended
$29/mo

The practical plan for an indie product with real users.

  • 20,000 presolved lookups
  • 500 real-time solves
  • 2,000 Postflop cache queries · per subscription service period
  • for indie devs
  • max 12G decision tree
    ?If a tree exceeds this limit, intelligent pruning may be applied or the solve may fail.
Choose plan

Pro

$99/mo

For production apps with recurring study and review traffic.

  • 100,000 presolved lookups
  • 3,000 real-time solves
  • 20,000 Postflop cache queries · per subscription service period
  • production apps
  • max 12G decision tree
    ?If a tree exceeds this limit, intelligent pruning may be applied or the solve may fail.
Choose plan

Need dedicated capacity? Scale starts at $299/mo with a dedicated solver pool and a max 72G decision tree. Contact sales.

Builder is recommended because it supports a meaningful product launch without jumping directly to production-scale capacity.

Billing contract

Three independent meters.

Counters are shared by all API keys on one account. Paid quotas follow the subscription service period; Free follows its existing monthly period.

Fast path

Presolved lookup

Preflop and presolved flop data return in milliseconds from the strategy library.

/v1/gto/preflop, /preflop/range, /flop/tree, /range, /flop/projected-range, /turn/projected-range
Compute path

Real-time solve

A solve is charged only when /v1/gto/solver triggers new solver work.

/v1/gto/solver — only when it triggers a new solve
Builder / Pro

Postflop cache queries

Builder includes 2,000 and Pro 20,000 postflop cache queries per subscription service period, separately from presolved lookups and real-time solves. An admitted root costs 1; every Turn or River card selection costs 1, including repeats. A cache miss or read failure still consumes an admitted request. Ordinary reads with a valid token are free. All API keys on an account share the counter. Coverage is not guaranteed.

/v1/gto/postflop/cache/query
Not chargedThese calls do not consume the corresponding counter.
Presolved detail/flop/node
Cached solver reuse/solver/tree, /solver/node, /evs, reusing a cached solve
Solver pool busyHTTP 429 status: busy · back off and retry.

Counters are shared by all API keys on one account. Paid quotas follow the subscription service period; Free follows its existing monthly period. Builder includes 2,000 and Pro 20,000 postflop cache queries per subscription service period, separately from presolved lookups and real-time solves. An admitted root costs 1; every Turn or River card selection costs 1, including repeats. A cache miss or read failure still consumes an admitted request. Ordinary reads with a valid token are free. All API keys on an account share the counter. Coverage is not guaranteed. quota_exceeded / solve_quota_exceeded / cache_quota_exceeded: HTTP 429. Cache quota details. . Pre-admission inapplicable or unsupported missing consumes 0; 503 cache_charge_unknown carries no quota and is not automatically retried. HTTP 402 cache_quota_exceeded rejects quota depletion. Cache EV/EQ endpoints are available in production since 2026-11-18. POST /v1/gto/postflop/cache/ev and /eq accept only a returned action token and concrete hole_cards. On /query, include_ev:true or include_eq:true returns 400 invalid_request before Auth validation, consumption or Native dispatch; omission/false remains ordinary query behavior. Each admitted independent EV/EQ request consumes 1 existing cache unit, including hit/join, and every admitted success carries that caller's cache_quota. Pre-admission inapplicable/unsupported missing consumes 0; after confirmation, failure, timeout or cancellation can still be charged. 503 cache_charge_unknown carries no quota and is not automatically retried; exhaustion returns 402 cache_quota_exceeded. The retained local-isolated capture is deployed:false, not a production three-street success or real-billing capture.

Plan ledger

Compare the complete contract.

Compare capacity and postflop cache access. Only active API Builder / Pro includes the cache API; Scale custom quotas do not include it.

PlanFreeStarterBuilder / recommendedProScale
Monthly price$0$3$29$99From $299
Presolved lookups101,00020,000100,000Custom
Real-time solves5255003,000Custom
Postflop cache queriesNot includedNot included2,00020,000Not included
Best forEvaluationSide projectsIndie productsProduction appsHigh-volume systems
API + SDKsIncludedIncludedIncludedIncludedIncluded
Solver capacitySharedSharedSharedSharedDedicated pool

Checkout

Choose the plan first. Pay in the dashboard.

Starter, Builder, and Pro are available in Dashboard with Card / Stripe checkout.

Available payment methods

Secure checkout opens after sign-in and plan selection.

Card / Stripe

Monthly subscription hosted by Stripe. Card details never enter Pokerai.

Pricing FAQ

Answers before checkout.

Can I try the real solver for free?

Yes. Free includes 5 real-time solves in addition to 10 presolved lookups.

What happens when quota runs out?

For presolved lookups and real-time solves: The API returns HTTP 429 with quota_exceeded (presolved) or solve_quota_exceeded (solve). On Free, upgrade in Dashboard to continue; paid plans start with Starter at $3/mo. Persistent postflop cache returns HTTP 429 with cache_quota_exceeded when its quota is exhausted. It is available only on active API Builder / Pro subscriptions, and its quota resets with their subscription service period. Starter and Scale do not include persistent postflop cache. Source-verified migration, pending deployment: POST /v1/gto/postflop/cache/ev and /eq: Each admitted independent EV/EQ request consumes 1 existing cache unit, including hit/join; every admitted successful response carries that caller's cache_quota. Pre-admission inapplicable or unsupported missing consumes 0; 503 cache_charge_unknown carries no quota and is not automatically retried. HTTP 402 cache_quota_exceeded rejects quota depletion. The existing /query root and Turn/River deal quota exhaustion remains HTTP 429.

Do busy requests consume quota?

No. HTTP 429 status: busy is not charged; back off and retry.

Do plans use different APIs?

Persistent postflop cache is available only on active API Builder and Pro plans. Other API features keep the same integration contract.

How is postflop cache usage counted?

Builder includes 2,000 and Pro 20,000 postflop cache queries per subscription service period, separately from presolved lookups and real-time solves. An admitted root costs 1; every Turn or River card selection costs 1, including repeats. A cache miss or read failure still consumes an admitted request. Ordinary reads with a valid token are free. All API keys on an account share the counter. Coverage is not guaranteed. Cache quota details. Source-verified migration, pending deployment: POST /v1/gto/postflop/cache/ev and /eq: POST /v1/gto/postflop/cache/ev and /eq: Each admitted independent EV/EQ request consumes 1 existing cache unit, including hit/join; every admitted successful response carries that caller's cache_quota. Pre-admission inapplicable or unsupported missing consumes 0; 503 cache_charge_unknown carries no quota and is not automatically retried. HTTP 402 cache_quota_exceeded rejects quota depletion.

First request

Start with $0 and a real solve.

Create one API key, run the quickstart, and decide from actual usage instead of a sales promise.

Get your API key