Playground
Run requests from your browser through the exact same gateway as your production traffic — nothing simulated, every run billed, recorded, and visible in the Request Explorer.
Running a request
- Open the Playground and paste an API key. The key stays in the tab's memory — it is never stored.
- Your key's available models load automatically; pick one.
- Optionally set a system prompt, add conversation messages, and adjust temperature or max tokens.
- Toggle Streaming to watch tokens arrive live, and press Run. Stop aborts an in-flight stream.
Inspecting the result
The response panel shows status, round-trip latency, token usage (for non-streamed requests), and the request ID with a direct View in Request Explorer link — the settled charge is there. Copy the raw response, the JSON, or a ready-to-run curl, Python, or JavaScript version of the exact request you just built.
Streamed token counts
Streamed responses do not include usage in the stream itself. The
authoritative counts and charge appear with the request in the
Request Explorer after the stream
completes.