BlackRiver AI
INITIALIZING
BlackRiver Labs · Local voice

Private voice synthesis, inside your browser.

Multilingual text-to-speech, zero-shot voice cloning and automatic Parakeet transcription. The complete stack is prepared at launch; text, voice and inference remain on this device.

WebGPULocal inference
ParakeetAuto transcript
No limitAuto chunking
OmniVoice StudioComplete local model stack
Core Decoder Voice encoder Parakeet ASR
Script0 characters
Voice source
Choose reference audio
20
1.0x
AUTO · UNLIMITED
Preparing generation…
Planning chunks…ETA calculating…
BlackRiver Labs · OmniVoice

Preparing the local voice stack

Models are kept in persistent browser cache after the first successful load; later visits reuse the cached weights instead of downloading them again.
0%
Checking WebGPU
Persistent browser cache · Hugging Face on first visit
1OmniVoice core2.3 GB · WebGPU
2Audio decoder86 MB
3Voice encoder654 MB
4Parakeet ASRMultilingual · INT8