SolvScout / SolvTune
Release: Production (GA)
Classify your ODE and rank solver arms — free lifetime after registration
SolvScout analyses an ODE system (RHS function or trajectory data) and produces a
ProblemProfile — stiffness class, dynamics type, noise regime, and dimension.
SolvTune (same package) ranks solver arms with confidence labels and recommended settings, using Resonix empirical evidence embedded in the native library.
Customers install a release wheel from the Resonix account portal — no repository access required. SolvTune recommendations are native-library lookups; SolvScout does not run solvers locally.
What it does
You have an ODE model and need to know how stiff it is and which integrator to try first. SolvScout probes your RHS (or a recorded trajectory) and returns a ProblemProfile. SolvTune — same package — ranks solver arms with confidence labels using Resonix empirical evidence inside the native library.
Install the release wheel from your Resonix account. SolvScout / SolvTune is free for life after registration; you still activate a machine-locked seat. It does not run solvers locally — use recommend / compare for shortlists, then wire SolvSRK or SciPy in your app.
When to use it
Before integrating SolvSRK or choosing among SciPy / external backends — classify once, rank arms, optionally mint a SolvSRK profile token via recommend_product().
See Examples for runnable code.
Who it is for
- Engineers choosing among SciPy, SolvSRK, and external ODE backends
- Teams that want an empirical shortlist before wiring SolvSRK profile tokens
- Analysts comparing corpus-backed arm tables without a local solver farm
What ships
- Python wheel (
solvscout) bundlinglibsolvscout - Public API:
classify,recommend,compare,recommend_product,recommend_jump,get_config - Example journeys in product
python/examples/(see Examples) - Free lifetime seat after registration + machine activation
Related products
Integrator: SolvSRK. Filter primitive: SolvFilter.