Voice engineloads the first time you press Generate (~90 MB, then cached)
Text to speech 0 / 20,000
Ctrl+Enter generates. Your text never leaves this device.
Takes 0
Every generation lands here: play it, download it, or drop it in the Studio.
Script 0 lines
Speakers show up here after Build lines.
Lines —
00:00.0 / 00:00.0
Write a script on the left and press Build lines.
Space plays · Enter in a line regenerates it · lines generate in the background while you keep writing
Voice library
Blend lab make a new voice
Mix up to three voices. The new voice keeps its own sound every time you use it.
Audio in
Drop audio or videoMP3 · WAV · M4A · MP4 · MOV · WEBM — up to 60 minutes
No audio yet.
Transcript —
Timestamps land here line by line. Click a time to hear that spot; edit any line.
Dub new voice, same timing
Transcribe something first (tab 4), then come back. Each line is re-spoken in the voice you pick and placed at the same time as the original.
Dub lines —
No dub yet.
Clone a voice engine: checking…
Cloning, 22 more languages and stem separation (removing the original voice in a dub) run on your own computer with the Billo Voice engine. Start it once, keep this tab open, and everything switches on by itself.
curl -O https://billysorrells.com/billotech/apps/voice/engine/billo_voice.py pip install chatterbox-tts demucs torchaudio python3 billo_voice.py serveMac with Apple Silicon, Windows/Linux with or without NVIDIA, 8 GB of memory or more. Models download on first use (English ~2 GB, other languages ~3 GB, stems ~80 MB).
none
Cloned takes
Takes in a cloned voice show up here and in Takes on the Speak tab.