Description
BrainForge backends are not excluded from either hybrid or uniform backend provider changes. This means that switching Default Backend in the settings modal to e.g. Gemini will also switch the BrainForge model providers to Gemini.
The UI will still show them as being locked on BrainForge, but logging confirms that this code iterates over all personas (including BrainForge) and changes the backend provider.
Relevant code is: routes/chat.py --> resolve_llm_clients
Steps to Reproduce
- add a logger statement showing the client of each persona to
resolve_llm_clients after the client is switched
- rebuild the docker container
- in the UI, navigate to the settings modal --> Models
- switch the Default Backend
- start a new chat and review the logs
Relevant Code
Relevant code is: routes/chat.py --> resolve_llm_clients
Other Notes
No response
Description
BrainForge backends are not excluded from either
hybridoruniformbackend provider changes. This means that switching Default Backend in the settings modal to e.g. Gemini will also switch the BrainForge model providers to Gemini.The UI will still show them as being locked on BrainForge, but logging confirms that this code iterates over all personas (including BrainForge) and changes the backend provider.
Relevant code is:
routes/chat.py-->resolve_llm_clientsSteps to Reproduce
resolve_llm_clientsafter the client is switchedRelevant Code
Relevant code is:
routes/chat.py-->resolve_llm_clientsOther Notes
No response