Core modules
Speech-to-Text
Transcribe speech locally with configurable models and decoding settings.
Text-to-Speech
Generate speech from text with voice presets and offline synthesis.
VAD & Diarization
Detect speech segments and identify speaker turns for long recordings.
Enhancement & separation
Clean noisy audio and separate overlapping voices using local models.
Enhancement
Reduce background noise and improve speech clarity on-device.
Separation
Split mixed speakers into individual tracks for analysis.
Privacy-first
Voice Lab keeps processing local by design. You control which files and models are used.
Ads (if enabled) are served via Google AdMob and follow regional consent requirements. See the privacy policy for details.