Point the camera at embossed or printed Braille, hear it read aloud, and ask questions about the page. Built accessibility-first.
Read Braille aloud, and ask about it.
A lightweight on-device capture loop feeds a dedicated Braille detector. A language model only cleans the result and answers questions — it never reads the dots itself.
A YOLO cell detector plus a classical dot-grid reader, auto-selected per image. ~99% cell accuracy on well-captured real Braille.
Spoken cues, earcons that rise as you align, and Core Haptics direct blind users to frame the page — then capture automatically.
Hear the page read aloud and ask questions by voice. A readability gate refuses to voice noise instead of guessing.
A photometric Braille generator — embossed dots, every angle, lighting, clutter — fine-tunes a YOLO11 detector on 15k labeled images.
Output routes through VoiceOver and refreshable Braille displays; adjustable speech rate; one-tap sighted mode.
A burst of full-resolution photos with per-cell consensus voting cancels noise from a single shaky frame.
Free, accessibility-first, and built for blind and low-vision readers. Star the repo to follow along — the iOS release is on the way.