Troubleshooting
Common errors, rate-limit messages, slow models, and other things that can go wrong (with fixes).
-
Why am I rate limited?
You've hit your tier's daily run cap or per-minute API rate. Wait for reset or upgrade. The error message tells you which limit + when it resets.
-
Tool returned an error, what now?
Read the error code, retry once, check your inputs. Most errors are transient — second attempt usually succeeds.
-
I lost my output, can I recover it?
If you saved it: yes, in /account/runs/. If you didn't save it: usually no — outputs aren't stored by default for privacy reasons.
-
Why is my model slow?
Default models run on shared CPU infrastructure (5–60s typical). Switch to a premium model for 1–5s response, or check /status/ for current load.
-
Why does the chat say it doesn't know?
The chatbot's knowledge has a training cutoff and is grounded only in our public docs + tools. For specifics about your account, use the support contact form.
