You are considering a second OpenAI-compatible provider. Before you commit, run these five checks against your own integration. Nothing you enter leaves your browser.

Each check has three answers: Yes, Partly, or No. Answer all five and the page shows a readiness band with the two guides that help you close any gap. Your answers stay on this page and are never sent anywhere.

1 Contract compatibility

Does your client work by changing only the base URL and API key?

Check the chat completions path, the Authorization header, and the streaming flag.

2 Request and response shape

Do the fields you actually use appear in the response?

Look for choices, usage, finish_reason, and the shape of streaming deltas.

3 Cost model

Can you compute cost per successful request from the usage object?

Separate input tokens from output tokens, then apply the provider's rates.

4 Limits and failure behavior

Do you know the rate limits and what your app does on 429 or a timeout?

Decide your retry and backoff policy, and what is safe to retry twice.

5 Data handling

Do you know where requests go and whether content is stored or used for training?

Answer all five checks to see your band. (0 of 5 answered.)