It is open source, and the whole thing runs on your own computer. What it does:
• Clones a voice from a 3 second sample (the README says 5 to 15 seconds sounds better)
• Designs a new voice from a text description
• Dubs video across languages
• Dictation and transcription
• Audiobook and long-form generation
16 text to speech engines, 11 speech recognition engines, a 646 language catalogue, and no account, API key, subscription or usage meter for the local workflow. 33,827 stars on GitHub on 21 September 2026.
It’s AGPL-3.0 and in active beta, so read the licence before you ship it to a client. Intel Macs can’t run the local backend.
#voicestudio #elevenlabs #voicecloning #opensource #texttospeech #ai #aitools #techupdates
![]()