Augmentative communication
Their world. Their words. Their voice.
EyeSay is a communicator-authored AAC application built on a simple architectural commitment: the language a communicator learns by location stays where they learned it. A stable motor plane of 48 positions holds vocabulary, and situational suggestions live in a separate adaptive plane that cannot move it.
Visit EyeSay AACeyesayaac.com(opens in a new tab)EyeSay is a communicator-authored AAC application built on a simple architectural commitment: the language a communicator learns by location stays where they learned it. A stable motor plane of 48 positions holds vocabulary, and situational suggestions live in a separate adaptive plane that cannot move it.
Self-advocacy is always reachable, partner modelling is clearly identified as the partner's, and only the communicator's own action makes the device speak.
The problem
Many AAC systems reorder or re-rank vocabulary in response to context. For a communicator who has learned words by motor position, a helpful re-ranking can erase the very stability that made the system usable.
Intended users
Nonspeaking and minimally speaking autistic children, their families and the professionals supporting them; extensible to other AAC users.
What EyeSay AAC does.
Voice Canvas
Forty-eight slots with permanent identifiers. Adding vocabulary uses an unused slot; it never shifts a learned one.
Context Halo
Eight situational suggestions in their own adaptive field, with no write path to the permanent grid — by design, not by discipline.
Agency Strip
STOP, NO, HELP, BREAK and TOILET remain directly reachable across the experience.
Model Mode
Partners can model language on the same topology without writing into the communicator's utterance.
Communicator-authored speech
Nothing is spoken until the communicator presses SPEAK.
Beyond requesting
Designed so a communicator can refuse, interrupt, comment, question, remember and joke — not only ask.
- Learned motor positions are structurally protected from re-ranking.
- Context adapts without destabilising language.
- Partner modelling is visibly distinct from communicator authorship.
- A word learned this morning is in the same place next term.
- Self-advocacy vocabulary is never more than one reach away.
- Support from adults without their words becoming the child's.
Two planes, one of which cannot move the other
The separation is enforced architecturally: the ranking that fills the Context Halo has no write path into the Voice Canvas, so adaptive support cannot reorganise learned language.
- 48 permanently identified slots on a stable motor plane.
- Separate 8-slot adaptive context plane with no write path to the grid.
- Always-reachable self-advocacy row.
- Partner modelling recorded separately from communicator utterances.
Private early access, currently delivered as an installable web application with a native iPad build in development. EyeSay is a communication application. It is not eye-tracking software, not a diagnostic or assessment tool, and not a treatment; no clinical efficacy is claimed, and it does not replace speech and language therapy input. Switch scanning, eye-gaze validation and specialist hardware compatibility are roadmap items and should not be assumed supported.
Visit EyeSay AACeyesayaac.com(opens in a new tab)