Text-to-Speech (Read Aloud)
RealTimeX can read text aloud using your configured text-to-speech provider. Playback starts progressively for longer content, so you can begin listening without waiting for the full response or note to finish generating.
Where you can use it
- Chat messages: Click the speaker button on a message to start playback.
- Personal Notes: Open a Markdown note and click the speaker button in the editor toolbar to read the current note aloud.
How playback works
- Long text is split into smaller chunks for smoother playback.
- Audio begins as chunks become available instead of waiting for the full text.
- Click the same control again to pause playback.
- You can use the same speech experience across chat and non-chat surfaces such as Personal Notes.
Supported providers
RealTimeX read-aloud works with configured local or cloud TTS providers, including:
- Local providers such as Piper and Supertonic
- Cloud providers such as OpenAI, Groq, ElevenLabs, and OpenAI-compatible services
Note: The shared read-aloud controls are not shown when your TTS provider is set to the native system/browser option.
Tips
- Use a headset or speakers with clear output if you are listening to long content.
- In Personal Notes, the current note content is read aloud directly from the editor.
- If a cloud provider is selected, confirm its API key and voice settings are configured in your environment.
Troubleshooting
- No speaker button appears
- Check that a supported TTS provider is configured instead of the native system/browser provider.
- Audio does not play
- Check your device volume, active output device, and whether the browser tab is muted.
- Playback fails with a provider error
- Verify the selected provider is configured correctly and any required API credentials are available.