Skip to main content

Create Custom Model

Create the empty model first; you'll add training data and run training in the next step (Training Process).

Open the Dialog​

On the Custom Models page, click New custom model in the header (or Create your first custom model in the empty state).

Fill the Form​

The dialog has two required fields:

  • Name - any label that helps you recognise the model later (e.g. Cardiology dictation EN).
  • Language - the base language. Only trainable languages appear in this dropdown (the panel filters by is_trainable: true).

The Create button stays disabled until both fields are filled and the create request isn't already in flight.

New custom model dialog Name + Language fields, Close and Create actions.

On success a Custom model created successfully toast appears and the model shows up in the table with Training status: Not running and no datasets yet.

If the server rejects the request, the error message is shown in a toast.

Open a Model​

Click the model's Name or Base language cell, or use the row's Edit action. This opens the model details page at /custom-models/<id>, with breadcrumbs back to the list.

The details page has four summary cards at the top:

  • Name - editable inline.
  • Training status - same badge as the list.
  • Base language model - full language name (e.g. English (US)).
  • Last modified - the last time anything changed on the model.

Right below the cards, an expandable Training status banner shows a help message tailored to the current status, plus a Need help preparing your data? section with a Force Alignment link. The banner auto-expands when the status is Success.

The detail page also auto-refreshes every 20 seconds.

Model details page top section Detail page header with breadcrumbs, the four summary cards, and the status banner.

Rename​

Edit the Name field inline and click Update in the page header. You'll get a Custom model updated successfully toast.

Only the name can be changed after creation - the base language is fixed.

Delete​

On the list, open the row menu → Delete. A destructive confirmation dialog opens; confirm and the model and its datasets are removed. You'll see a Custom model deleted successfully toast.

Delete confirmation dialog Delete confirmation - destructive, cannot be undone.

Next Steps​