Skip to content

Latest commit

 

History

History
55 lines (33 loc) · 1.44 KB

README.md

File metadata and controls

55 lines (33 loc) · 1.44 KB
Raw

vacuum-voice-gen

Generate the voice of your rooted Dreame vacuum robot using ChatGPT and ElevenLabs.

Requirements

  • A Linux environment.
  • Bun.

Usage

Set the environment variables

Copy .env.dist to .env and fill in the environment variables.

Edit the ChatGPT prompt

The voice lines are generated using the personality of a character described in a ChatGPT prompt, which is located in data/prompt.txt. Sorry, for my use I've written this in French but feel free to translate it back to your native language before editing it.

Generating the audio files

bun install
bun run src/generate.ts

The output files are in data/voices and data/voice_pack.tar.gz.

Importing the audio files

Please refer to https://github.com/Findus23/voice_pack_dreame#installation.

Example of voices

You can import an example of a French voice pack I made for my personal use.

Archive : https://github.com/Ilshidur/vacuum-voice-gen/releases/download/1.0.0/voice_pack.tar.gz md5 : c86c7dc23fb2432ba9ed21e391904197 And just call it BASTIANO.

TODO

I'll add a Docker image later.

Credits

Thanks a lot to Findus23 who maintains https://github.com/Findus23/voice_pack_dreame, which I heavily copied.

Don't forget to 🌟 Star 🌟 the repo if you like this project !
Your feedback is appreciated