Settings is the ongoing configuration area after first-run setup. The settings layout provides breadcrumbs, a scrollable main region, and a small footer.
Overview Groups
The settings overview groups navigation cards into:
| Group | Entries |
|---|---|
| General | General, System, Providers, Channels |
| Capabilities | Model settings |
| AI Features | Skills, MCP |
| Automation | Heartbeat, Cron, Task Manager, Apps |
Each overview card navigates to a route from the settings navigation constants.
Provider Settings
Route: /settings/providers
Provider settings use the same provider catalog as the start page. The page loads saved API key status, shows masked saved keys, and lets the user connect or edit provider credentials.
Saving uses window.app.setProviderApiKey(providerId, draft).
Model Settings
Model setting cards are listed in the settings navigation.
| Capability | Route id | UI status |
|---|---|---|
| Assistant | friday |
Runtime-backed provider/model settings. |
| Speech to text | speech-to-text |
Runtime-backed provider/transcription model settings. |
| Text to speech | text-to-speech |
Read-only pending configuration with ElevenLabs / Rachel - multilingual. |
| Text to image | image-assistant |
Runtime-backed settings page, but model catalog is placeholder unless a provider exposes image models. |
| Text to video | text-to-video |
Read-only pending configuration with video-provider-coming-soon. |
| Text to audio | music-creator |
Read-only pending configuration with music-provider-coming-soon. |
The shared detail page uses one reusable implementation. Assistant, speech-to-text, and image creator are runtime-backed and can save provider/model choices. Text-to-speech, video, and audio currently show pending configuration notices with disabled provider/model selectors.
Assistant-Specific Settings
The assistant detail page includes:
- Chat history link
- Provider selector
- Model selector
- Reasoning effort selector for OpenAI and DeepSeek providers
- Save action
The save action stores the selected provider/model through the preload API.
Speech-To-Text Settings
Speech-to-text filters providers to those with available transcription models. The save action stores the selected provider/transcription model through the preload API.
The current explicit speech-to-text model catalog is OpenAI
gpt-realtime-whisper.
Future UI Coverage
The docs and store model include additional module settings for text-to-video, text-to-audio/music, OCR, and embedding. Renderer settings currently expose pending settings pages for video and audio. A dedicated embedding settings page is not present in the current renderer navigation.