ElevenLabs
ElevenLabs voices speak your radio calls, voicemails and announcements. It is a cloud service: the text of each voice line is sent to ElevenLabs and billed to your ElevenLabs account. The app shows the request and asks for your approval first, and replaces names, phone numbers, emails and addresses with placeholders by default: see What goes to cloud models. Nothing about your documents is sent, only the lines to be spoken (and the sound note, if you switch sound effects on).
1. Get an API key
Section titled “1. Get an API key”Create a key in your ElevenLabs account under Developers → API keys. Speech uses the credits of your plan.
2. Add the provider
Section titled “2. Add the provider”- Providers → Add provider → Type: ElevenLabs. The address is fixed.
- Paste the key into API key. Leave Default model empty to use
eleven_multilingual_v2, or pick another speech model from the list. - Add provider, then Test on the card. The test only checks the key and counts your voices. It generates no speech and uses no credits.
- In Routing, assign Audio to ElevenLabs. You can pick a different model on that row.
3. Choose voices
Section titled “3. Choose voices”Open Edit on the provider. The Voices table has one row for each role in the open project, plus Any other speaker. Choose a voice for each from the voices on your account. A role with no voice of its own uses the one chosen for Any other speaker. If that is also empty, the audio step stops and asks you to pick a voice.
The roles show only while a project is open. Choices you make for a role stay saved when you close the project.
Audio quality
Section titled “Audio quality”ElevenLabs sends audio at 44.1 kHz by default, which its Pro plan and above include. If you see “Your ElevenLabs plan does not allow this audio format”, set Audio quality to 24 kHz in the provider settings.
Sound effects (optional)
Section titled “Sound effects (optional)”When Add sound effects when an inject has a sound note is ticked, an inject whose media brief has a sound note (for example “crowd murmur, PA chime”) also gets a short effect track, mixed quietly under the voice. It is off by default, sends the note to ElevenLabs and uses extra credits.
Messages you may see
Section titled “Messages you may see”| Message | What to do |
|---|---|
| ElevenLabs rejected the API key | Paste the key again in Edit. |
| Your ElevenLabs credits are used up | Add credits or wait for your plan to reset. |
| ElevenLabs does not know the voice chosen for … | The voice was removed from your account. Pick another in Voices. |
| ElevenLabs is rate limiting requests | Wait a minute and run the step again. |
| ElevenLabs refused … | The text breaks the service’s content rules. Reword the line. |
Audio from one inject is limited to 50 MB.
