Ollama returning 400 error code
✅Solved
I have openclaw connected to local Ollama and recently tried to change hte model. Now when I send requests, I see in Ollama logs it is returning 400 to openclaw. Howeber openclaw doesn't log what the 400 response is. I have console logs and verbose logging but can not seem to get any output on what the response is/why it's 400ing. I am trying to use ollama/translategemma:12b.
Solution
Yes — but only as a “text-only” model.
OpenClaw assumes it can send a tool schema (
Two implications:
1) Auto-discovery will hide it (OpenClaw filters Ollama models to tool-capable ones).
2) If you force it, you must disable tools for that model (then it can’t browse/write files/run exec/etc while using it).
OpenClaw assumes it can send a tool schema (
toolstool_choicetoolsTwo implications:
1) Auto-discovery will hide it (OpenClaw filters Ollama models to tool-capable ones).
2) If you force it, you must disable tools for that model (then it can’t browse/write files/run exec/etc while using it).
Config pattern that usually fixes the 400
Add the model explicitly and deny tools for that provider/model:
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
83,815Members
Resources
Recent Announcements
Similar Threads
Was this page helpful?
