Voice and music with AI

You no longer need a studio to work with sound. There are three jobs, each with its own tool.

1. Text to speech (voice-over)

ElevenLabs turns written text into a natural voice. Handy for videos, ads, audiobooks and training material.

How:

1. Paste your text.
2. Choose a voice (male/female, age, tone).
3. Listen, and adjust punctuation if needed — commas and full stops control the pauses.

You can clone a voice too, but only your own or someone’s who agreed.

2. Music and songs

Suno and Udio produce a full song from a few lines.

Upbeat pop jingle for a coffee shop, warm female vocals,
acoustic guitar and light percussion, 30 seconds

You can write the lyrics yourself, even in Uzbek or Russian. Pronunciation isn’t always perfect, so generate several versions.

3. Speech to text

Recorded a meeting, interview or lecture? The Whisper model turns speech into text and supports Uzbek and Russian. Many apps and services are built on it.

Once you have the text, give it to Claude or ChatGPT: “Write a short summary of the meeting and a list of action items.”

Legal note

• Check the service’s licence before using AI music commercially.
• Cloning someone else’s voice without consent may be illegal.

Exercise

Create a 20-second jingle for your business or channel and add a short announcement voiced with ElevenLabs.