aaPanel_Kern Hi,
I am trying to use a custom OpenAI-compatible API endpoint with aaPanel AI Assistant (self-hosted proxy).
I successfully connected my endpoint and:
/v1/models works correctly/v1/chat/completions returns valid OpenAI responses- requests reach my server without issues
However, the AI Assistant UI does not display any response, even though the API returns valid data.
Also, I encountered issues with unsupported parameters like:
enable_searchenable_thinkingstream_options
which I had to strip out manually in my proxy.
My question is:
👉 What exact response format does aaPanel AI Assistant expect from /v1/chat/completions?
Does it require a custom wrapper (e.g. msg, data.content, etc.) instead of standard OpenAI format?
Is there any documentation for the expected API schema?
Right now it seems the assistant is tightly coupled with your internal endpoint (/aap_agent/openai/v1) and not fully compatible with custom OpenAI APIs.
Thanks in advance.
