common-voice/common-voice

Common Voice is part of Mozilla's initiative to help teach machines how real people speak.

What it solves

Common Voice is a platform designed to solve the scarcity of open, high-quality speech data. It allows people to donate their voice recordings of specific sentences to create public domain datasets for training voice recognition tools.

How it works

The project is a web application that facilitates the collection of speech donations. Users contribute by reading sentences provided by the platform, and the system manages the process of collection, translation, and language-specific workflows to build a massively multilingual speech corpus.

Who it’s for

It is for developers and researchers who need open-source speech datasets for training AI voice recognition models, as well as volunteers who wish to contribute to the open web.

Highlights

  • own a massively-multilingual speech corpus
  • public domain datasets (CC0 license)
  • community-driven collection of voice recordings
  • integrated workflows for adding new languages and sentences

Related

  • Project
  • Project
  • Project
  • Project
  • Project