Available now

Eleven v4: text to speech

By ElevenLabs

Eleven v4 reads up to 5000 characters in a voice you name, and audio tags in the text direct how each line is delivered.

Model ID
Maker
ElevenLabs
Added
September 29, 2026

Input

Fields marked * are required. Review the price before generating.

The text to convert to speech. Supports audio tags such as [whispering] and IPA pronunciation enclosed in forward slashes.

Example. Edit it or write your own.

Additional settings

The voice to use for speech generation

Use a suggested voice or enter an exact voice ID.

Voice stability. Lower values allow more expressive delivery; higher values make delivery more consistent.

How closely the output follows the reference voice. Higher values increase similarity but may reduce naturalness.

Seed for best-effort reproducibility. Identical output is not guaranteed.

Language code (ISO 639-1) for speech generation and text normalization.

Whether to normalize text such as numbers and dates before generation.

Whether to return character-level timing information with the generated audio.

Price$0.13 / run

You sign up first. Your settings are kept.

About this model

Overview

Eleven v4 turns text into speech. Put an audio tag such as [whispering] before a line to change its delivery, and write a word's IPA pronunciation between forward slashes when it must be said a certain way.

Stability trades expressive delivery for an even one, and similarity sets how closely the read follows the voice. Choose MP3, PCM, Opus or telephone formats, and turn on character timing for subtitles or lip sync.

What you can build

  • Acted voice-overs with whispers and shifts in tone
  • Audiobook chapters with directed delivery
  • Brand and place names said exactly right
  • Phone prompts in telephone audio formats

Prompting tips

  • Put the audio tag right before the words it should change: "[whispering] Don't wake them."
  • Lower stability for a livelier read, raise it to keep a long script even.
  • Write tricky names in IPA between slashes instead of respelling them.

What to check

Listen for pronunciation, pauses and any cue read aloud. Check names and numbers against the script.

Inputs

  • Text (required): The text to convert to speech. Supports audio tags such as [whispering] and IPA pronunciation enclosed in forward slashes.
  • Voice (optional): The voice to use for speech generation.
  • Stability (optional): Voice stability. Lower values allow more expressive delivery; higher values make delivery more consistent. From 0 to 1.
  • Similarity Boost (optional): How closely the output follows the reference voice. Higher values increase similarity but may reduce naturalness. From 0 to 1.
  • Seed (optional): Seed for best-effort reproducibility. Identical output is not guaranteed. From 0 to 4294967295.
  • Language Code (optional): Language code (ISO 639-1) for speech generation and text normalization.
  • Output Format (optional): Output format of the generated audio. Options: Mp3 22050 32, Mp3 44100 32, Mp3 44100 64, Mp3 44100 96, Mp3 44100 128, Mp3 44100 192, Pcm 8000, Pcm 16000, Pcm 22050, Pcm 24000, Pcm 44100, Pcm 48000, Ulaw 8000, Alaw 8000, Opus 48000 32, Opus 48000 64, Opus 48000 96, Opus 48000 128, Opus 48000 192.
  • Apply Text Normalization (optional): Whether to normalize text such as numbers and dates before generation. Options: auto, on, off.
  • Timestamps (optional): Whether to return character-level timing information with the generated audio.

Output and limits

  • File formats: MP3, PCM, μ-law, A-law, Opus.
  • Sung or spoken words can come out unclear or differ from your text.
  • Exact length, tempo and key are not guaranteed; listen to the whole result before you use it.
  • Prompts that name real artists or protected songs can be refused.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • up to · 1,000 ASCII characters: $0.13 / run