Skip to content

Find or add your first model

A model in DeltaLLM points to an AI provider and one of its models. Your application calls the public name you choose, while DeltaLLM keeps the provider details and credentials separate.

Video tutorial: Add a model and call it. Watch on YouTube.

Find the sample model

The Docker starter configuration can add a model named gpt-4o-mini. Get the list of available models:

curl http://localhost:4002/v1/models \
  -H "Authorization: Bearer YOUR_MASTER_KEY"

Replace YOUR_MASTER_KEY with the DELTALLM_MASTER_KEY value from your .env file.

If the response includes gpt-4o-mini, your model is ready. Continue to Send your first request.

Add a model in the Admin UI

Use these steps if the model list is empty or you want to use a different provider:

  1. Open http://localhost:4002 in your browser.
  2. Sign in with the administrator email and password from your .env file.
  3. Open Gateway.
  4. Open Models.
  5. Select Add model.
  6. Enter a public name. Use gpt-4o-mini to use the examples in this guide.
  7. Select the provider.
  8. Enter the provider's model name.
  9. Add the provider API key or select a saved provider credential.
  10. Keep the type set to Chat for a standard text model.
  11. Save the model.
  12. Wait until the model shows a healthy status.

Run the model-list command again. The response should include the public name you chose.

Continue to Send your first request.

If the model does not appear

  • Check that the provider API key is present and valid.
  • Check that the provider model name is correct.
  • Review the container logs for a connection or sign-in error.
  • See Models in the Admin UI for more help.

For every available setting, see the model deployment reference.