Captions
The Captions widget displays live auto-generated captions on your overlay, transcribing your voice in real time. It improves accessibility and helps viewers watch with sound off.
How It Works
Captions are generated from your microphone audio using speech-to-text. Text appears on the overlay a few seconds after you speak and fades out after a configurable delay.
Setup
- Open Captions in the dashboard.
- Grant microphone access when prompted.
- Select your input device (the mic you speak into, not the stream output).
- Choose your language — captions work best when the language matches your speech.
- Add the Captions widget to the Compositor and position it (usually bottom-center of screen).
Settings
| Setting | Description |
|---|---|
| Language | The language you're speaking. |
| Font size | Caption text size. |
| Display duration | How long each caption line stays visible (seconds). |
| Max lines | How many lines of captions to show at once. |
| Background | Optional semi-transparent background behind caption text. |
Language Support
Captions support the same languages as the dashboard: English, Spanish, Japanese, Russian, and Korean, plus additional languages depending on your browser's speech recognition support.
Notes
- Captions require microphone access in the browser. Grant permission when asked.
- Caption accuracy depends on your microphone quality and speaking clarity.
- Captions are local — they're generated in your browser and not stored or sent to any server.