Documentation

What SolverForge solves, when to use it, and how it differs from mathematical solvers.

Current release, published package status, runtime surface, and roadmap.

Start with the CLI shell, then continue into a complete hospital, lessons, deliveries, or field-service use case.

Install the solverforge PyPI package, write Python models with decorators and callbacks, and run the native SolverForge engine.

Scalar assignment, retained jobs, schedule constraints, and browser updates in one worked app.

Lesson timetabling with teachers, cohorts, rooms, two scalar variables, retained jobs, and timetable views.

List-variable vehicle routing with maps, retained snapshots, and insertion recommendations.

Field-service routing with technicians, skills, shifts, road-network travel, and route geometry.

Generated accessors, source-aware streams, joins, projections, collections, and scoring terminals.

solver.toml, phase configuration, move selectors, acceptors, foragers, and facade exports.

Retained scoring-only rows from projections, joined pairs, and directed self-joins.

Retained solve lifecycle, snapshots, events, analysis, pause, resume, cancel, and delete.

Road-network loading, routing, matrices, route geometry, and map-backed planning helpers.

Frontend assets, lifecycle adapters, retained-job controls, and scheduling views.

Benchmark harness, CVRP and employee-scheduling runs, TOML configuration, and PostgreSQL result storage.

Which crate or companion repository owns each part of the runtime surface.