Playground
Send a governed call from the console and see what it cost, how long it took, and the record it left, without writing code.
The Playground is not a sandbox. Calls go through the same auth, budget checks, metering and audit as your production traffic.
Send a call
Pick a model
The Model field takes any model your workspace can reach.
mock-gptis a built-in test model that costs almost nothing. Use it to test the setup, not output quality.Load an agent, optionally
Load from agent pulls in an existing agent's model and system prompt.
Write the prompt
A system message is optional; the user message is the call.
Generate
Leave streaming on to watch tokens arrive. The budget check runs before the first token either way.
Read the receipt
Each completion is followed by a receipt:
| Field | What it tells you |
|---|---|
| est. cost | What this call cost, priced from the token counts. |
| model | Which model answered. If routing or a fallback stepped in, this can differ from the one you asked for. |
| prompt / completion / total | Token counts. |
| TTFT | Time to first token. |
| latency | Total time for the call. |
| Audit · Trace | Links to this call's audit entry and trace. |
The cost on a streamed completion is an estimate. The billable figure is on the Budgets and Audit ledgers. They usually agree; if they don't, the ledgers are correct.
Compare two models
The Compare tab runs the same prompt against two models side by side, with a receipt under each. Use it to check whether a more expensive model is better for your prompt. Both calls are billed.
Watching the budget
The Monthly budget meter in the top right updates as you spend. Budget alerts you configured fire here too.
Over the cap, you get the same 402 your application would. See
what happens when it is refused.