ElevenLabs
ElevenLabs turns a text into spoken audio using the voices in your ElevenLabs account, including any custom voices you have created there. The result is an audio file you can send, store or play from other components.
When to use it
- Turn an assistant’s answers into audio to send them as voice notes.
- Generate the voice-over for a video or a presentation from a script.
- Create audio versions of newsletters or internal announcements.
- Read daily summaries aloud for people who prefer to listen to them.
Configuration
The component needs an Eleven Labs API credential with your ElevenLabs key. See Credentials.
| Field | What it does |
|---|---|
| Credential (required) | The ElevenLabs credential. |
| Text (required) | The text to turn into audio. You can type it or connect the output of another component, such as an agent’s answer. |
| Voice (required) | The voice that reads the text. The list shows the voices in your account with their gender, language and accent; use the refresh button to load it. |
| Model (required) | The ElevenLabs voice model used. |
| Custom Voice ID | Optional. The identifier of an ElevenLabs voice. When filled in, this voice is used instead of the one chosen in Voice. |
Output
The Audio Base64 output returns the generated audio, in MP3 format, encoded as text so it can travel through the flow. Connect it to components that send or store audio files in that form.
Example
A training school wants to send its students an audio summary of each class:
- An agent summarizes the class notes into a short text.
- Connect the agent’s answer to Text on ElevenLabs and pick the school’s voice in Voice.
- Send the audio from Audio Base64 to the students through the channel you use.
Good practices
- Write the text the way it would be said aloud: short sentences and no abbreviations the voice cannot read.
- ElevenLabs usage depends on the amount of text converted; check your account’s plan if you will generate a lot of audio.