AIAI Actions
Text-to-Speech
Convert text to spoken audio messages
Generate voice messages from text using AI text-to-speech synthesis.
How to Use
- Type your message in the editor
- Open AI Actions → Text-to-Speech
- Select a voice and model
- Adjust speed if needed
- Preview the generated audio
- Send as a voice message
Options
| Option | Description | Default |
|---|---|---|
| Voice | Select from available voices (see below) | Alloy |
| Speed | Playback speed from 0.25× to 4× | 1× |
| Model | TTS-1 (fast), TTS-1 HD (high quality), or GPT-4o Mini TTS | TTS-1 |
The model resolves in order: your AI Preferences setting → the assistant's configured TTS model → TTS-1 (default).
Available Voices
| Voice | TTS-1 | TTS-1 HD | GPT-4o Mini TTS |
|---|---|---|---|
| Alloy | ✓ | ✓ | ✓ |
| Ash | ✓ | ✓ | ✓ |
| Coral | ✓ | ✓ | ✓ |
| Echo | ✓ | ✓ | ✓ |
| Fable | ✓ | ✓ | ✓ |
| Nova | ✓ | ✓ | ✓ |
| Onyx | ✓ | ✓ | ✓ |
| Sage | ✓ | ✓ | ✓ |
| Shimmer | ✓ | ✓ | ✓ |
| Ballad | — | — | ✓ |
| Verse | — | — | ✓ |
| Marin | — | — | ✓ |
| Cedar | — | — | ✓ |
Ballad, Verse, Marin, and Cedar are only available with the GPT-4o Mini TTS model. Selecting them with TTS-1 or TTS-1 HD will not work.
Preview
Before sending, you can:
- Listen to the generated audio
- Regenerate with a different voice
- Adjust speed and regenerate
- Cancel and type manually instead
Requirements
- OpenAI only — OpenAI provider must be connected
- Not available when using Anthropic as the AI provider
For TTS in automation flows, see the AI Text-to-Speech Card.