Run7 AIGet an API key ↗
RUN7 / LEARN & BUILD

AI API checker: key connectivity, model calls and streaming

By Run7 AI · Updated: 2026-11-17

Enter an authorized endpoint and test key for one minimal request. The test runs in your browser and checks this connection, not model identity or future uptime.

Advanced: output-limit field and API version

Select according to your endpoint reference. Output is capped at 32; a reasoning model may exhaust the allowance without visible text, making the result inconclusive.

Test result

No request sent.

What does this tool check?

TestWhat success means
Minimal textAuthentication and a text response shape worked for this model.
StreamingText fragments and a termination marker arrived. Timing includes local network and preflight.
Model listA list was retrieved; not proof of invocation permission for each model.

Fixed short prompt, no automatic retries, one request per run, 60-second timeout and 1 MiB response limit. No load testing, paid fingerprinting or authenticity certificate. Cancellation and timeout cannot guarantee that service processing or billing stopped.

What if the browser blocks the response?

The endpoint must allow cross-origin requests from this page. Messages direct access also uses the browser-access header where supported. Without CORS permission the browser may hide the HTTP status; this is not proof of an invalid key. Keep browser security enabled and use the local cURL template.

Open local cURL and setup instructions · Troubleshoot 401 / 429 and timeouts

How are keys and results handled?

The checker uses no browser storage, cookies or analytics and uploads no key, endpoint or response body to Run7. The key exists briefly in page memory and is cleared after a request or on leaving. Only a diagnostic summary is displayed; raw responses are not displayed or saved. Your device, browser extensions and chosen endpoint can still access requests. Use a low-credit test key. The Run7 host may still record ordinary page-access logs.