Unlikely. What is your use case for that? Would it be better to use a http client for that purpose?
Aaron Kunaseelan
•
Dec 17, 2025
It would be useful to pass flags and specify parameters without needing you to create a custom knob for each provider.
For example, OpenAI has xhigh reasoning and "search_context_size": "high".
Aaron Kunaseelan
•
Dec 18, 2025
cool, you already implemented the search context!
I just imagine it being hard to keep up
Daniel Nguyen
Team•
Dec 18, 2025
Yeah you can find the search context size in the Plugins > OpenAI. I’m adding the new xhigh reasoning.
The reason I don’t want to support arbitrary json config is it’s going to be a maintenance nightmare. I might add that option but will need to be careful + probably behind a feature flag with no guaranteed of future compatibility.
Log in to comment and vote
Comments5
Daniel Nguyen
Dec 17, 2025
Unlikely. What is your use case for that? Would it be better to use a http client for that purpose?
Aaron Kunaseelan
Dec 17, 2025
It would be useful to pass flags and specify parameters without needing you to create a custom knob for each provider.
For example, OpenAI has xhigh reasoning and "search_context_size": "high".
Aaron Kunaseelan
Dec 18, 2025
cool, you already implemented the search context!
I just imagine it being hard to keep up
Daniel Nguyen
Dec 18, 2025
Yeah you can find the search context size in the Plugins > OpenAI. I’m adding the new xhigh reasoning.
The reason I don’t want to support arbitrary json config is it’s going to be a maintenance nightmare. I might add that option but will need to be careful + probably behind a feature flag with no guaranteed of future compatibility.
Aaron Kunaseelan
Dec 18, 2025
Thanks, I understand.