Two ways to generate voice
| Built-in voices | Your own API keys | |
|---|---|---|
| Setup | Built in | Add your GenAiPro or Minimax key in the dashboard |
| Cost | Uses your plan's credits | You pay the provider directly |
| Languages | 30+ | Depends on the provider |
| Best for | Getting started, multilingual channels | Creators who already use a provider's premium or cloned voices |
The built-in options include standard text-to-speech voices and ElevenLabs voices.
Features
- Voice controls: adjust speed and pitch, and stability for ElevenLabs voices.
- Standalone or chained: voice only, or script → translation → voice → images → render (rendering on VIP and Lifetime).
- Long scripts: narration for videos up to 2 hours.
Tips for long-form narration (10+ minutes)
- Write for the ear: short sentences, and numbers written as they're spoken ("twenty twenty-six").
- Split by chapter: generate per section so you can fix one part without regenerating everything.
- Lock one voice per channel: viewers notice when the narrator changes.
- Check names and terms: listen to the first minute of every chapter before you render.
- Using your own or a cloned voice? YouTube asks creators to disclose realistic altered or synthetic content. Its disclosure help notes that using AI to clone your own voice for a voiceover doesn't require disclosure. See Is AI content allowed on YouTube?
Use cases
Documentaries, history and story channels, educational explainers, multilingual versions of existing scripts (translation and TTS), and audiobook-style narration. Monetizing a channel that uses AI voices? Read can you monetize YouTube videos with an AI voice?
FAQ
Which languages are supported?
30+ languages with the built-in voices. With your own keys, your provider's language list applies.
Can I use ElevenLabs voices?
Yes. ElevenLabs voices are available as a built-in option. You can also add your own GenAiPro or Minimax API key.
Do I need to disclose AI voices on YouTube?
Disclosure is about realistic content that viewers could mistake for real. Narration of a clearly produced explainer is a different case from a synthetic voice impersonating a real person. Read YouTube's guidance in our policy guide.