ElevenLabs icon

ElevenLabs

WIP

Speech → Text to Speech

AI-generated

Overview

The ElevenLabs node's Speech - Text to Speech operation converts input text into spoken audio using customizable voice models. This node is useful for automating the generation of natural-sounding speech from text, enabling applications such as:

  • Creating voiceovers for videos or presentations.
  • Generating audio content for accessibility (e.g., reading articles aloud).
  • Producing dynamic audio responses in chatbots or virtual assistants.
  • Testing and prototyping voice user interfaces.

Users provide the text and select a voice, then receive an audio file output that can be played, saved, or further processed.

Properties

Name Meaning
Text The text string to convert into speech. Required. Example: "Be good to people!"
Voice ID The voice to use for speech synthesis. Can be selected from a searchable list of available voices or specified by entering a voice ID directly.
Additional Fields Optional advanced settings to customize the speech output:
  • Binary Name Change the name of the binary output data. Default: "data".
  • File Name Change the output audio file name. Default: "voice".
  • Streaming Latency Enable latency optimizations at some cost to audio quality. Integer between 0 and 4. Default: 0 (off).
  • Output Format Audio format of the generated speech file. Default: "mp3_44100_128".
  • Model Name or ID Identifier of the voice model to use. Can be chosen from a list or specified via expression.
  • Stability Controls voice stability (how consistent the voice sounds). Number between 0 and 1. Default: 0.5.
  • Similarity Boost Adjusts how closely the voice matches the original speaker. Number between 0 and 1. Default: 0.75.
  • Style Exaggerates the voice style. Number between 0 and 1. Default: 0.
  • Speaker Boost Boolean flag to activate speaker boost feature. Default: false.
  • Seed Fixed seed number for reproducible results. Default: 0.
  • Stitching Enables stitching mode to give the model context by passing past and previous text. Boolean. Default: true.

Output

The node outputs a single item per input with the following structure:

  • json: Contains metadata about the generated speech (not detailed in the source but typically includes status or identifiers).
  • binary: Contains the synthesized audio file data under a key named by the Binary Name property (default "data"). This binary data represents the audio in the specified Output Format (default MP3).

This allows downstream nodes to play, save, or manipulate the audio file.

Dependencies

  • Requires an API key credential for ElevenLabs API access.
  • Makes HTTP POST requests to https://api.elevenlabs.io/v1.
  • Uses ElevenLabs API endpoints for listing voices, models, and performing text-to-speech synthesis.
  • No other external dependencies are indicated.

Troubleshooting

  • Invalid or missing API key: The node requires a valid API key credential. Ensure it is configured correctly in n8n credentials.
  • Voice ID not found: Selecting a voice ID that does not exist or is unavailable will cause errors. Use the provided searchable list or verify the ID.
  • Unsupported output format: Specifying an invalid output format may cause the API to reject the request.
  • Network issues: Connectivity problems to the ElevenLabs API endpoint will result in request failures.
  • Parameter validation errors: Values outside allowed ranges (e.g., stability > 1) may cause errors.
  • Beta status: As the node is marked BETA and under active development, users might encounter unexpected bugs or incomplete features.

Links and References

Discussion