Non-invasive Speech Restoration
Giving Voice to
Those Who've Lost Theirs
VOCL is a non-invasive EMG headset that decodes attempted speech from facial muscle signals, restoring natural communication for people with ALS, aphasia, and dysarthria.
The Problem
7.5 million Americans live without the ability to speak.
Their brains work perfectly, but failing muscles trap them in silence. Current solutions fail them: AAC tablets cost $10,500 and force users to type at 10-15 words per minute—ten times slower than natural speech. Brain implants offer better results but require invasive surgery and cost $50,000-$100,000, excluding 99% of patients.
The VOCL Process
User Attempts Speech
User thinks about speaking or attempts to vocalize
EMG Signal Capture
8 electrodes detect microvolt signals from facial muscles
Signal Processing
Filtering, ICA, and PCA-LDA reduce noise and extract features
AI Decoding
CNN-LSTM model classifies phonemes and predicts words
LLM Enhancement
Context-aware error correction produces natural output
Speech Output
Text or synthesized speech delivered in under 300ms