Settings
Settings Overview
All 10 settings tabs — shortcuts, input, microphone, web navigation, and smart detection.
Access Settings via:
- Sidebar: Click the Settings gear icon
- Keyboard:
Cmd + ,(macOS) /Ctrl + ,(Windows)
Vocoding Settings are organized into 10 tabs:
| # | Tab | Purpose |
|---|---|---|
| 1 | Settings | Shortcuts, input behavior, web navigation, smart detection |
| 2 | Transcription | Whisper models, language, VAD, streaming |
| 3 | Local LLM | Ollama local configuration |
| 4 | LLM | Groq, OpenRouter, Ollama Cloud |
| 5 | Automation | Auto-copy, auto-paste, multi-task queue, Smart Bubble |
| 6 | Agents | Agent management (also available as full view) |
| 7 | Profiles | Optimization profile management |
| 8 | Storage | Retention policies and cleanup |
| 9 | Appearance | Theme, opacity, font size, language |
| 10 | Library | Saved prompts (also available as full view) |
Note: The Library tab is hidden when you are already in the Library view (
Cmd + 5).
Keyboard Shortcuts
Configure global hotkeys that work from any application.
| Shortcut | Description | macOS Default | Windows Default |
|---|---|---|---|
| Chat / Dictate | Main recording hotkey | Option + Space | Alt + Space |
| Transcribe (Raw) | Transcribe without LLM | Cmd + Shift + R | Ctrl + Shift + R |
| Optimize (LLM) | Transcribe with optimization | Cmd + Shift + O | Ctrl + Shift + O |
| Context Injection | Capture clipboard context | Cmd + Shift + Space | Ctrl + Shift + K |
To change a shortcut:
- Click on the shortcut field
- Press your desired key combination
- The shortcut saves automatically
Requirements: Must include at least one modifier key (Cmd, Option, Ctrl, Shift). Cannot conflict with system shortcuts.
Input Behavior
| Setting | Description | Default |
|---|---|---|
| Context Injection | Enable/disable context injection shortcut | ON |
| Enter to Send | Send message on Enter (Shift+Enter for new line) | ON |
| Auto-save Chat to Library | Automatically save chat messages to Library | OFF |
Microphone
| Setting | Description |
|---|---|
| Device selector | Choose which microphone to use |
| Input level indicator | Visual meter showing audio input strength |
| Device health | Status indicator for the selected microphone |
Translate to English
| Setting | Description | Default |
|---|---|---|
| Translate to English | Convert non-English speech to English | OFF |
Note: Translation support varies by Whisper model. Some models (like large-v3-turbo) may not support translation.
Web Navigation
Enable AI agents to search the web and fetch content during conversations.
| Setting | Description | Default |
|---|---|---|
| Online Navigation | Master toggle for web search/fetch | OFF |
| Privacy-First | Prefer privacy-respecting search providers | OFF |
Search Providers (8+ available)
| Provider | Type | API Key Required |
|---|---|---|
| Brave Search | Core | Yes |
| Tavily | Core | Yes |
| Perplexity | Premium | Yes |
| Exa | Premium | Yes |
| You.com | Premium | Yes |
| SearXNG | Privacy | No (self-hosted URL) |
| Mojeek | Privacy | No |
| DuckDuckGo | Fallback | No (built-in) |
Each provider card shows:
- Status badge (active/inactive)
- API key input field
- Save/Update/Delete buttons
- Quota display (when applicable)
Smart Detection
Configure how Vocoding automatically selects agents and profiles.
| Setting | Description | Default |
|---|---|---|
| Smart Detection | Enable/Disable | Shows status: Smart / Deterministic / Off |
| Scope | What to auto-detect | All (agents + profiles) |
Advanced Settings (expandable section)
| Setting | Description | Default |
|---|---|---|
| Agentic enabled | Use AI for detection | ON |
| Provider mode | Auto or manual provider selection | Auto |
| Provider | Which LLM to use (if manual) | ollama |
| Model | Specific model for detection | (auto) |
| Timeout | Max time for detection (ms) | 3000 |
| Auto-Apply Confidence | Minimum confidence to auto-apply | 0.82 |
| Hard Abstain Confidence | Below this, never auto-apply | 0.55 |