All tutorials
UtilitiesPartial

Models

/models

TL;DR

Registry of integrated models and single-tool run surfaces.

Use it / Skip it

Use the model registry when you need to see what tools exist or run one model without entering a workflow-specific page.

Use when

Looking for a specific backend, checking model capabilities, or running a one-off model task.

Don't use for

Do not use it when a focused product page exists for the workflow; use Interface, Fold, Affinity, or RFantibody for supported workflows.

Inputs

Model key
The selected model from the registry.
Task payload
Model-specific input parameters.

Outputs

Run result
Model-specific output and job metadata.
Capability context
Registry description and category.

Walkthrough

  1. 1. Search or browse by model category

    Search or browse by model category.

  2. 2. Open the model card and review what it expects

    Open the model card and review what it expects.

  3. 3. Run a small one-off task

    Run a small one-off task.

  4. 4. Move successful workflows to the focused module for repeated use

    Move successful workflows to the focused module for repeated use.

Under the hood

  • Uses the platform model registry and /api/v1/models/{key}/run style endpoints.
  • Capability and readiness vary by model.

Worked example

Check whether a fold backend is available
Open Models, find the backend, review its card, then use Fold if the task is simply sequence-to-PDB.

Pitfalls

  • Registry presence does not mean a model is calibrated for your biological decision.
  • One-off model runs may lack the richer exports of product pages.

See also