-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Labels
enhancementNew feature or requestNew feature or request
Description
Summary
Integrate claude-code-router to enable using multiple AI model providers (OpenAI, Google, Z.AI) alongside Anthropic models. Panopticon will manage the router configuration and provide a settings UI.
Installation
pan install
- Install claude-code-router as part of the standard installation process
- Configure it to work with Claude Code
pan sync
- Check if claude-code-router is installed, install if missing
- Sync any model configuration changes
Settings Page
Add a new settings page (or section) in the dashboard with:
1. Specialist Agent Models
Configure which model each specialist uses:
- Review Agent: [model dropdown]
- Test Agent: [model dropdown]
- Merge Agent: [model dropdown]
2. Planning Agent Model
- Planning Agent: [model dropdown]
3. Task Complexity Models (ties into PAN-75)
Configure models based on task difficulty:
- Low complexity tasks: [model dropdown]
- Medium complexity tasks: [model dropdown]
- High complexity tasks: [model dropdown]
4. API Keys Management
Secure storage and management of API keys for providers:
- OpenAI API Key
- Google AI API Key
- Z.AI API Key
- (Anthropic key already configured via Claude Code)
Default Model Options
Include these models out of the box:
Anthropic (via Claude Code)
- claude-opus-4
- claude-sonnet-4
- claude-haiku-3.5
OpenAI
- gpt-4o
- gpt-4o-mini
- o1
- o1-mini
- gemini-2.0-flash
- gemini-1.5-pro
Z.AI
- glm-4-plus
- glm-4-flash
Technical Notes
- Panopticon owns the claude-code-router configuration (users don't edit it directly)
- Store settings in
~/.panopticon/settings.jsonor similar - API keys should be stored securely (consider using system keychain or encrypted file)
- Settings changes should take effect without restarting agents
Acceptance Criteria
-
pan installinstalls claude-code-router -
pan syncchecks/installs claude-code-router if missing - Dashboard has settings page for model configuration
- Can configure model per specialist agent
- Can configure planning agent model
- Can configure models for task complexity levels
- Can add/edit/remove API keys for providers
- Model dropdowns only show models for which API keys are configured
- Configuration persists across restarts
- Changes apply to newly spawned agents
Related
- PAN-75 (Task complexity-based model routing)
Metadata
Metadata
Assignees
Labels
enhancementNew feature or requestNew feature or request