Native macOS poker study workspace

MelaGTO

A complete study and solving workspace for Mac. MelaGTO combines bundled preflop exploration, custom postflop tree construction, local MelaSolver GPU and CPU computation, strategy and EV inspection, imported hand review, and decision training in one native interface.

GPU + CPUlocal solver backends

13solver endpoint semantics

5hand-history site families

Updated July 27, 2026 · Maintained by Pokerai API

MelaGTO / macOSreal product interface

MelaGTO home screen with Preflop Explorer, Custom Solution, Import Hand, GTO Trainer, and recent games

MelaGTO strategy browser with action frequencies and a range grid

01 / system map

From a hand or a blank canvas to a local solved strategy.

MelaGTO keeps study data, interface state, and numerical computation in clear layers. The SwiftUI app gathers a spot from a preflop chart, a custom setup, or an imported hand. A sandboxed XPC service carries the validated configuration into the local solver core, then returns progress, strategy, EV, and tree data to the workspace.

01 / source

Explore or import

Start from bundled preflop ranges, a saved workspace, or a supported hand-history file.

02 / model

Define the tree

Confirm positions, ranges, board, stack, bet sizes, raise caps, all-ins, and accuracy.

03 / compute

Sandboxed XPC

A stable JSON contract transports solver work without coupling the SwiftUI shell to C++ internals.

04 / engine

MelaSolver

Run the GPU or CPU backend locally, then query the completed tree for strategy and EV.

02 / complete product tour

Eight real screens. One connected study workflow.

The product is not a thin solver launcher. Each surface owns a distinct part of study: finding a known spot, creating a new one, monitoring computation, navigating a solved tree, practicing decisions, or turning a real hand into a reproducible solve.

Workspace home

Choose the shortest path into study

Continue recent work or start in Preflop Explorer, Custom Solution, Import Hand, or GTO Trainer.

MelaGTO home with four study workflows and recent games

Preflop explorer

Compare complete mixed ranges

Navigate versions, positions, and action contexts while viewing side-by-side 13×13 frequencies.

GTO trainer

Practice against a solved decision

Play the hand, choose an action, and compare it with the strategy and leak analysis for that spot.

MelaGTO trainer table with decision analysis

Custom solution

Build a spot before spending compute

Set the stack, format, positions, ranges, board, and postflop configuration in a guided workflow.

Local solve

See the workload and live progress

Validate the configuration, run MelaSolver GPU or CPU, monitor iterations and memory, stop safely, or release resources.

Strategy browser

Inspect actions, EV, nodes, and runouts

Move through the solved tree, switch between strategy and EV, inspect hand classes, and choose later-street cards.

MelaGTO strategy browser with action frequencies

Tree builder

Control the abstraction explicitly

Configure per-street bet, raise, and donk sizes, all-in thresholds, raise caps, and accuracy before solving.

Import hand

Turn played hands into reviewable solve drafts

Detect supported sites, store imports locally, expose warnings, inspect the action timeline, and create a Custom Solution draft.

03 / capability ledger

A complete desktop workflow, not a collection of disconnected tools.

ExplorePreflop Browser

Browse bundled chart versions by position, situation, and action context; carry selected ranges into a custom setup.

ConfigureCustom Solution

Validate board, ranges, stack, action tree, accuracy, iteration limit, isomorphism, and memory constraints before submission.

ComputeMelaSolver GPU / CPU

Run the full postflop solve locally through a sandboxed XPC service with progress, stop, release, and resource reporting.

InspectStrategy + EV Browser

Navigate action nodes and runouts; inspect mixed frequencies, combo detail, EV grids, and later-street branches.

ReviewImported Hands

Parse supported cash-game histories, deduplicate locally, surface structured warnings, and convert eligible heads-up postflop hands into solve drafts.

PracticeGTO Trainer

Replay decisions in a table interface, compare the chosen action with solved frequencies, and review recurring deviations.

OrganizeQueue + Saved Solves

Manage longer study sessions, preserve solver output as compressed local archives, and reopen completed trees offline.

ExportCSV + PNG

Export readable strategy and EV data or share the current grid as an image without rebuilding the spot elsewhere.

04 / implementation model

Native UI around a stable solver boundary.

The app does not expose C++ engine ownership to SwiftUI. Typed Swift models become a JSON request, the XPC service dispatches one of the established solver operations, and the workspace observes structured progress and results. The boundary also lets MelaSolver CPU and GPU share the same product behavior.

01SwiftUI owns navigation, editing, validation, and presentation.

02A sandboxed XPC process isolates long-running solver work.

03A stable C ABI shields the app from engine implementation details.

04Solver computation remains local; imported hand data is stored locally.

local solve contractproduct-level flow

workspace
  .validate(spot, ranges, tree, accuracy)
  .initialize(resourceDirectory)
  .submitSolve(
    backend: "MelaSolver GPU",
    forceReplace: true
  )

while task.status == "solving" {
  workspace.update(
    iteration: task.currentIteration,
    memory: task.physicalFootprint
  )
}

strategy = workspace.query(
  node: "root/CHECK/BET",
  runout: ["9h"]
)

workspace.export(strategy, formats: [
  "csv",
  "png"
])

05 / study loop

From a real question to a reusable answer.

01 / Start

Find or import the spot

Begin with preflop charts, a recent workspace, a blank custom solution, or an eligible imported hand.

02 / Define

Confirm ranges and tree

Review positions, board, effective stack, action sizes, raise limits, all-ins, and convergence target.

03 / Solve

Compute on the Mac

Run MelaSolver GPU or CPU, watch progress and memory, and keep the numerical workload outside the UI process.

04 / Study

Navigate and retain

Inspect strategy and EV, practice the decision, export a grid, or save the tree for later offline review.

Scope and provenance

MelaGTO is a macOS product maintained by Pokerai API. It is a first-party product record, not a third-party customer case study. This page documents the local study workflows visible in the application: range exploration, custom solving, hand review, and practice. The postflop solver runs on the user's Mac through MelaSolver GPU or CPU. The hand-history importer currently targets supported cash-game formats from PokerStars, GGPoker, 888/Pacific, Winning Poker Network, and Bovada/Ignition; unsupported variants are rejected with structured reasons rather than guessed. This page makes no claim about customer adoption, downloads, benchmark performance, commercial partnerships, or external endorsements. MelaGTO is for poker study, analysis, and practice. It does not enable real-money betting, deposits, withdrawals, payouts, cash games, gambling transactions, or real-time assistance at live real-money tables.

Build a connected product

Put strategy inside the workflow.

Use Pokerai API when your product needs presolved lookups or real-time solving without shipping solver infrastructure.

Get your API key →