Record the voiceovers for your videos with your own cloned voice, no microphone or studio

With ten seconds of your voice, a free local program generates the voiceovers for your product videos, announcements and tutorials. You type the script, your voice comes out. Nothing is uploaded to the internet.

Level
No code
Done from claude.ai or ChatGPT, without programming.
Time
Half an hour for the first voiceover
Published

What it costs

Zero. The application is open source and runs on your computer; it does not charge per character or per minute. A free alternative to subscription voice services.

guided mode // I ask, you answer

ready to start

Do it with me, step by step

I will ask you 7 questions about your business, use your answers to get the prompt ready to copy, and then walk you through the 5 steps, one at a time. Like a Thermomix: you just answer and move on.

Also from Claude or ChatGPT with the Master Cookbook MCP: say "start the recipe clonar-tu-voz-para-locuciones-de-videos".

Who it is for

Entrepreneurs and small marketing teams that make product videos, reels or tutorials and have no time, microphone or patience to record twenty takes.

The problem

Recording a clean voiceover takes silence, a microphone, time and repeating until it comes out right. Artificial voice services charge per character and sound like anyone but you.

What changes

You write the script, pick your cloned voice and in seconds you have the audio ready to lay over the video. According to the README, Voicebox clones voices from a short sample, generates in 23 languages and runs entirely on your machine.

Ingredients

  • A computer running macOS, Windows or Linux (a graphics card helps, but it is not required)
  • Ten seconds of your voice recorded with your phone in a quiet place
  • The video script written out, with the pauses marked

Step by step

  1. 1

    Record your voice sample

    Ten to twenty seconds reading something naturally, somewhere without noise. Speak the way you speak in your videos, not like you are reading a contract.

  2. 2

    Install the app and create your voice profile

    Download the installer for your system from voicebox.sh. Create a voice profile with the sample you recorded. The first time it downloads a model; after that it works offline.

  3. 3

    Paste the script and generate

    Write the voiceover text with clear punctuation: periods and commas mark the pauses. Generate and listen. If a sentence sounds odd, rewrite it more simply.

  4. 4

    Export and lay it over the video

    Export the audio and bring it into your usual video editor (CapCut, iMovie, Premiere). Lower the music volume so the voice stays in front.

  5. 5

    Build a library of phrases

    Save the closing lines, greetings and calls to action you repeat in every video. Next time you only generate the new part.

The prompt

Replace what is in brackets with your company details, or use the guided mode above: it asks you and fills it in for you.

# Role
You are the voiceover scriptwriter for [BRAND NAME], which sells [WHAT IT SELLS] to [TARGET AUDIENCE]. You write to be read aloud by a cloned voice: short sentences, natural rhythm, no words that are hard to pronounce.

# Context
The video lasts [VIDEO LENGTH IN SECONDS] seconds and shows [WHAT THE VIDEO SHOWS]. The brand tone is [TONE, FOR EXAMPLE "warm and direct"]. The final call to action is [CALL TO ACTION, FOR EXAMPLE "message us on WhatsApp"].

# Task
Write the voiceover script for that video, with the number of words that fits the length (about 2.5 words per second), marking pauses with periods and commas, and without acronyms or long numbers that sound bad when read aloud.

# Output format
The script ready to paste into the voice generator, in a single block, and below it an alternative version 20% shorter.

# Examples
15-second video for a bakery: "Birthday this week? Handmade cakes, baked the same day. Order today on WhatsApp and it is at your door tomorrow."

What usually goes wrong

  • Recording the sample with background noise or in a put-on announcer voice. The clone copies what you give it: give it your real voice, in silence.
  • Scripts with acronyms, long figures and thirty-word sentences. They sound bad in any voice; write the way you talk.
  • Cloning other people's voices. Only your own, or with written permission.

When not to use this recipe

Only your own voice, or that of someone who gave you written consent: cloning a third party's voice is a legal and ethical problem. For a large campaign with art direction, a professional voice actor is still better.

Source and real case

Marketing · video content

The repository, trending weekly on GitHub at the time of review, describes a local, open source voice studio that clones voices from a short sample, generates in 23 languages with several engines, dictates into any app and exposes an API and MCP. The recipe uses it to solve voiceovers for a small business's videos at no cost and without leaving the machine.

Trending repository:jamiepine/voicebox — The open-source AI voice studio · GitHub

Last verified:Repository README read on 22 September 2026 (cloning from a few seconds of audio, 23 languages, seven voice engines, all local, API and MCP server); the URL returned 200.

VoiceVideoVoiceoverContent

Share it

If someone on your team could use this recipe, send it in one click. The suggested text copies itself.

Want to apply it in your business? Message Cristián directly.

Telegram