Agent Lab · AI Gateway status
Is it the Gateway,
or is it us?
This page asks our hosted AI Gateway whether it is ready, straight from your browser, with no sign-in. It asks again every minute.
AI Gateway (api.agentlab.cresclab.com)
Checking…
No answer yet
What each state means
What you see, and what to do
Operational
The Gateway answered and said it is ready to take model calls.
What to do If your call still fails, the cause is most likely that request, its key or a limit: look up the code in the response in Gateway error codes, or open the request under AI Gateway › Usage.
Not serving right now
The Gateway answered but said it cannot take model calls yet, for example while it starts or reloads its configuration. Calls in the meantime can get a 503
gateway_unavailable.What to do This is on our side. Wait and retry; the OpenAI and Anthropic SDKs retry on their own. If it stays like this for more than a few minutes, write to info@cresclab.com with the time you saw it.
Unreachable
Your browser got no answer from the Gateway within 10 seconds. The Gateway may be down, or your network, company firewall or proxy may be blocking api.agentlab.cresclab.com.
What to do Open this page from another network, such as a phone hotspot. If it works there, ask your IT team to allow
api.agentlab.cresclab.comon port 443. If it is unreachable from every network, it is most likely on our side: write to info@cresclab.com with the time.
What this page checks
The Gateway itself, nothing more
- This page only asks whether our hosted Gateway is ready. It does not tell you whether a model provider has capacity, or whether your workspace has access or room in its limits.
- The answer is what this device on this network sees. It is not stored, and there is no history.
- To check from a server:
curl https://api.agentlab.cresclab.com/v1/gateway/readiness. An answer of{"status":"ready"}means it is operational. For a self-hosted Gateway, ask your own server's/v1/gateway/readiness.