Text to Speech Block

Convert text into spoken audio using ElevenLabs voices. Returns base64-encoded MP3 audio. Authenticate with an ElevenLabs API key.

Overview

When to Use

• Generate spoken audio narration from AI-written text in an automated workflow
• Build voice response systems that read back dynamic content to users
• Produce audio files for notifications, summaries, or announcements without a human recording session
• Convert long-form content (articles, reports) into podcast-style MP3 audio
• Power multilingual TTS pipelines by pairing with a translation block upstream

Configuration

Operation

Selects the action the block performs. There is currently one option:

Text

The text content to convert to speech. Accepts multi-line input and supports {{blockName.field}} references to pass in text generated by earlier blocks. Visible only when Operation is tts_synthesize.

• Type: long-input
• Required: yes (for synthesis)
• Placeholder: Hello, welcome to our service!

Voice ID

The ElevenLabs voice identifier that determines the speaker's voice. Visible only when Operation is tts_synthesize.

• Type: short-input
• Required: yes (for synthesis)
• Placeholder: 21m00Tcm4TlvDq8ikWAM

Find voice IDs in the ElevenLabs Voice Library or via the ElevenLabs API.

Model

The ElevenLabs model to use for synthesis. Defaults to eleven_turbo_v2_5 when left blank. Visible only when Operation is tts_synthesize.

• Type: short-input
• Required: no
• Placeholder: eleven_turbo_v2_5

Common model IDs include eleven_turbo_v2_5 (fast, low-latency) and eleven_multilingual_v2 (higher quality, multilingual).

ElevenLabs API Key

Your ElevenLabs API key used to authenticate every request. Store it as a workflow secret and reference it as {{ELEVENLABS_API_KEY}} rather than pasting the raw value.

• Type: short-input (password-masked)
• Required: yes

Inputs & Outputs

Inputs:

• operation (string) — Operation to perform; must be tts_synthesize
• apiKey (string) — ElevenLabs API key
• text (string) — Text to synthesize
• voiceId (string) — Voice ID
• modelId (string) — Model ID

Outputs:

• data (json) — Base64 audio with MIME type and format; shape: { audioBase64: string, mimeType: "audio/mpeg", format: "mp3" }
• metadata (json) — Response metadata; shape: { voiceId: string, characterCount: number }

Tools

Text to Speech Synthesize (tts_synthesize) — Calls the ElevenLabs /v1/text-to-speech/{voiceId} REST endpoint with the supplied text and model, streams back the MP3 binary, and returns it as a base64-encoded string along with synthesis metadata (voice ID used and character count).

YAML Example

tts_1:
  type: tts
  name: "Text to Speech"
  inputs:
    operation: tts_synthesize
    apiKey: "{{ELEVENLABS_API_KEY}}"
    text: "{{summarizer.content}}"
    voiceId: "21m00Tcm4TlvDq8ikWAM"
    modelId: "eleven_turbo_v2_5"
  connections:
    outgoing:
      - target: next-block-id