-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Description
Description
Modify the agent to allow changing LLM models during a session while preserving conversation context.
Requirements
- Agent tracks current model selection
/modelcommand triggers model switch- Conversation history preserved across model changes
- Banner/UI shows current model name
- Graceful handling if API key for selected provider is missing
Implementation
- Add
currentModelfield to Agent struct - Create model registry with provider mappings
- Update
processTurn()to use current model's client - Show model name in UI/banner
Related Issues
- Implement /model slash command for switching LLM models #18 - /model command
- Abstract LLM client interface for multi-provider support #19 - LLM client abstraction
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels