WAYS TO USE HELIX OPTIMIZATION

From a browser tab
to your own process.

Use a hosted planning experience directly, add optimization to your product through an API, or bring execution into your infrastructure. Choose who owns the interface and where the solve runs.

01 / For planners and domain teams

Use the hosted workspace

Work directly in a browser: select a supported problem, prepare inputs, configure the manifest-backed controls, run a solve and review the plan. Start without building your own user interface.

Explore example workspaces →
02 / For product and integration teams

Call the hosted API

Keep your own user experience. Your backend sends structured problem records and configuration to a provisioned solver endpoint, then brings the result and diagnostics back into your application.

Read the service guide →
03 / For teams controlling infrastructure

Operate a private service

Keep the API boundary while placing the solver service and workers in agreed infrastructure. Multiple applications can share a solver deployment without embedding the runtime in each process.

Plan a private deployment →
04 / For teams owning the runtime

Embed the Python solver

Call a supported Python solver entry point from your own process. Build the solver into an application, batch job or internal planning tool with application-owned execution.

Read the embedded guide →
COMPARE THE EXPERIENCE

Choose your starting point.

The configuration and result contract remain explicit across the choices. Available engines and problem families depend on your deployment.

Your browser → application backend → solver service

Use the hosted workspace

Your team owns the problem data, configuration and acceptance of the plan. Hosting and enabled features are agreed during onboarding.

Request a hosted walkthrough. Workspace access and available problem families are provisioned; this site is not a live solver console.

Explore example workspaces →
ONE AUTHORING APPROACH

Configure once. Choose the execution boundary.

People or agents can draft versioned manifests against the supported schema. Validate the configuration against the target deployment before using it. A change of hosting model does not guarantee identical engines, dependencies or supported features.

Configure with an agent →

See what the planner sees.

Explore routing and rostering interactions based on our existing applications.

View UX examples ↗