Full access is free during Beta. A paid subscription will be offered after Beta.

ElevenLabs — User Guide

Ultra-realistic TTS and voice cloning—many languages.

Visit website
Freemium
Strengths
  • The naturalness of the voice is the highest in the industry, with rich emotional expression and almost indistinguishable from real people.
  • Sound Clone: ​​Clone any sound with just 1 minute of audio
  • Supports 29 languages, with consistent dubbing quality in multiple languages
  • Voice Design: Creating new sounds from text descriptions
  • The API is complete and can be integrated into various applications such as podcasts, audiobooks, and customer service systems.
Best for
  • Podcast and audiobook voiceovers (no real person recording required)
  • YouTube video narration and course dubbing
  • Product demonstration video narration
  • Multilingual content localization (generating multiple language versions of the same content)
  • Virtual assistant and customer service voice system

Basic text-to-speech

The most basic and commonly used function of ElevenLabs is to convert text into natural speech.

Scenario

Generate narration for YouTube videos

Prompt example
In the ElevenLabs text-to-speech interface:

1. Select the voice: Rachel (English, professional female voice) or Adam (English, male voice)
2. Adjust parameters:
   - Stability: 0.5 (balances stability and expressiveness)
   - Clarity: 0.75 (sharpness)
   - Style: 0.3 (moderately stylized)
3. Enter text:
"Welcome to today's tutorial on AI tools. In this video, we'll explore
how to use ElevenLabs to create professional voiceovers in minutes."
4. Click Generate
Output / what to expect

Generate high-quality English narration audio:

  • Natural intonation and reasonable pauses
  • Moderate emotions and strong sense of professionalism
  • Downloadable MP3 format for direct use in videos
  • Generation time is about 5-10 seconds
Tips

Lower Stability is more expressive but may be unstable, higher is more stable but may be monotonous. For narration content, Stability 0.4-0.6 works best.

Scenario

Generate Chinese dubbing

Prompt example
Select a Chinese voice (such as "Xiaoxiao" or upload a Chinese voice clone) and enter:

"Welcome to today's tutorial. Today we will learn how to use artificial intelligence tools
Improve work efficiency. First, let’s take a look at the most basic features. "

Speech speed setting: 0.9 (slightly slower, suitable for tutorials)
Output / what to expect

Generate natural Mandarin Chinese dubbing:

  • Accurate pronunciation and correct intonation
  • The speaking speed is moderate and suitable for the teaching content
  • Can be used directly in course videos or product demonstrations
Tips

It is recommended to use special Chinese voices for Chinese content to avoid using English voices to generate Chinese (accent problem).

Starter & above

The rest of this guide

Additional scenarios and the full comparison table are included with Starter and above. Sign in with an eligible account to load them.

You're on the Free plan. Upgrade to Starter or higher to unlock the rest of this guide—additional scenarios and the full comparison table.

Loading full guide…