Status
Live numbers from our own request log, refreshed every minute. Failed upstream requests are never billed to you.
Status unavailable
Last updated —
Could not reach the status endpoint.
Daily request volume and success rate. Days with no traffic are shown as gaps rather than as 100%.
Success rate counts every request that reached a model and returned a full response. Interrupted streams count as failures even when partially billed. Requests rejected before reaching a model — bad key, blocked model, no balance — never enter the log, so they neither help nor hurt the rate. Latency is the median end-to-end time including upstream generation. First-token times average streamed responses only.
Create an account, grab a key, and make your first call in minutes.