> For the complete documentation index, see [llms.txt](https://docs.agenticflow.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.agenticflow.ai/reference/nodes/elevenlabs_group_speech.md).

# ElevenLabs: Speech and Transcription

Generate speech or multi-voice dialogue, change a voice, transcribe audio, align words with audio, or manage speech generations.

Find this node under **ElevenLabs** in the action picker. It uses an [ElevenLabs connection](/integrations/connection-providers/media/elevenlabs.md).

## Set up the node

1. Add **ElevenLabs: Speech and Transcription** to your workflow.
2. Select your ElevenLabs connection.
3. Choose **Model / Action**. The configuration form changes to match your choice.
4. Fill the required fields, using values from earlier steps where needed.
5. Save your changes and run the step. Review its result in **Build** or **History**.

## Parameters and inputs

The fields depend on the selected action. Use the labels, required markers, and available choices shown in your form.

| Input           | What to provide                                                                                               |
| --------------- | ------------------------------------------------------------------------------------------------------------- |
| Model / Action  | Choose the speech, dialogue, transcription, alignment, or speech-engine action you need.                      |
| Text or audio   | Enter a script for speech; provide an audio file for transcription or voice conversion.                       |
| Voice and model | Choose a voice and compatible model when required. For dialogue, configure the individual lines and speakers. |
| Existing item   | For a retrieval or management action, provide the relevant transcript, engine, or generation ID.              |

See [model-specific forms](/workflows/media-action-forms.md) for dropdowns, media uploads, repeatable lists, and using earlier results.

## Outputs

An audio file, transcript, timing information, or speech-generation details, depending on the selected action.

Changing from speech generation to transcription clears a previously selected speech model. Select a model suitable for the new action before running.

## Example

Choose **Text To Speech**, select a voice and a speech model, and enter a short product introduction. To transcribe a recording instead, switch to **Speech To Text**, choose a transcription model, and supply the recording.

<details>

<summary>Available Model / Action choices</summary>

* Text To Speech
* Text To Speech With Timestamps
* Text To Dialogue (Multi-Voice)
* Text To Dialogue With Timestamps
* Speech To Speech
* Speech To Text
* Get Transcript By Id
* Delete Transcript By Id
* Create Forced Alignment
* List Speech Engines
* Create Speech Engine
* Get Speech Engine
* Update Speech Engine
* Delete Speech Engine
* Create Speech Generation
* List Speech Generations
* Get Speech Generation

</details>

Availability of individual models and resources depends on your provider account. A successful connection does not grant access to every action.

## Related guides

* [ElevenLabs connection and action overview](/integrations/connection-providers/media/elevenlabs.md)
* [Test a step and reuse earlier results](/workflows/workflow-single-run/run-step.md)
* [Retrieve media and save it to AI Drive](/workflows/workflow-single-run/media-results.md)