Voice, music and avatars with AI

What you'll see in this lesson

1. ElevenLabs: voices that act, not just read

   Eleven v4, released on September 28, 2026, replaces v3 as the recommended model and understands cues like [laughs].

2. Cloning a voice: instant or professional

   Instant cloning copies a voice from less than 2 minutes of audio; professional cloning trains a model, but only with your own voice.

3. Free voices on your computer: Kokoro and Piper

   Kokoro and Piper generate speech offline with no per-character cost; they sound less expressive, but work for announcements, tutorials or drafts.

4. Music with Suno v6: who owns the song

   Suno v6 came out on September 9, 2026, built with Warner and BMG. Without a subscription you can't download songs or use them commercially.

5. Alternatives to Suno for music

   For ads and brand videos you have ElevenLabs music with a commercial license, and ACE-Step if you want it free and local.

6. Talking avatars: HeyGen and Hedra

   HeyGen creates presenters and clones you on video; Hedra animates a character from a photo and an audio clip.

7. Consent: in writing and for specific uses

   Using someone's voice or image for commercial purposes without their permission is illegal in Spain, even if the tool lets you.

8. What you can't do

   Cloning celebrities, imitating someone to deceive, or selling what you made on a free plan are the most common red lines.

9. Summary

   Voices, music and avatars already sound professional; what makes the difference is choosing a licensed plan and having consent.