/v1 API routes as your production code.
Open playground
Test image and video generation from the browser.
What you can test
API key handling
The playground requires an API key for requests.- The key is typed manually.
- The key must be locked before sending.
- The key is never persisted to storage.
- The key is cleared when the tab closes or reloads.
- Control characters and bidirectional override characters are rejected.
Region selection
Choose the same region as your API key:
Requests fail authentication if the key and regional hostname do not match.
Request editor
The form updates when you switch models. Supported fields can include:generation_promptgeneration_ratiogeneration_output_formatgeneration_output_numbergeneration_input_filegeneration_input_file_last_contentgeneration_durationgeneration_resolutiongeneration_generate_audio- model-specific fields
Send a generation
1
Select image or video
Open the tab that matches the model type.
2
Choose a model
Search by AI lab or model identifier.
3
Enter and lock your key
Paste a
bye_... key for the same region and lock it.4
Fill in request fields
Use the generated form to send only fields supported by the selected model.
5
Send the request
The playground calls the corresponding
/v1/generate/... route and shows
the response JSON.Track results
After a request starts, the response includes ageneration_id. The playground tracks the generation with realtime updates and falls back to polling if the realtime channel disconnects.
The response panel truncates very large JSON for display, but copied payloads include the full response subject to browser clipboard limits.
Manage a generation
Use the management section with a generation UUID:
Generation IDs must be valid UUIDs.
