tacit

tacit
Option Board
Technical Implementation

Inspiration

Patients with ALS, paralysis, stroke-related speech loss, or severe motor impairments may understand everything around them but still struggle to communicate.

We built Tacit so they can communicate using intentional blinks.

What it does

Tacit is a blink-driven communication system for patients who cannot reliably speak or use their hands.

Yes / No with blinks
AI-suggested responses
Blink-scanning keyboard
Text-to-speech
Patient-specific history

How we built it

Frontend: React + TypeScript inside Electron; simple patient and clinician interfaces built for fast, low-effort communication.

Blink Detection: Presage SmartSpectra detects blinks from the webcam, with per-user calibration and filtering for accidental blinks.

Communication Flow: Automatic scanning cycles through Yes/No, suggested responses, and keyboard options; an intentional blink selects the highlighted option.

AI Suggestions: Gemini uses patient context and previous interactions to generate likely responses and clinician questions.

Blink Keyboard: Row-and-column scanning lets patients build custom messages when a suggested response is not enough.

Text-to-Speech: ElevenLabs converts selected or typed responses into speech.

Storage: SQLite stores patient profiles, clinical context, sessions, responses, and communication history.