The person is not a diagnosis
Profiles set reply length, tile style and vocabulary. They never infer a diagnosis, an intelligence or a mental age. An adult can use single symbols.
Bridging disability and everyday AI.
The same wave of AI that writes everyone else's email has mostly passed AAC by. The reason is not indifference; it is that a communication system cannot afford the one thing generative models do most easily, which is to make something up.
A guess is not a kindness.
When an AI invents an email sentence, you edit it. When it invents a sentence for someone who cannot easily correct it, it has spoken for them. So the runtime refuses rather than improvises: a reply that does not fit the person's profile is discarded, not truncated into a different meaning; a failed translation is reported, not guessed at; manual composition is always available.
This is the same rule as the rest of Perslis
Peel, our science runtime, answers “unknown” when the evidence is not there. Accessibility is where that discipline came from: context and dependable expression decide whether a person is understood.
Four commitments you can check in the code.
The user's words are the user's
Only delivered selections and the user's own edits enter the model's context. Suggestions and unspoken drafts do not. Explicit edits survive the word limit.
Offline is a right, not a tier
The offline model is the one that launches by default on a configured machine. No feature is held back for a connected plan.
Say what is not finished
The adult checkpoint is evaluated, not released. The child model's response quality is not solved. Both are stated in the documentation, not discovered later.
Builders, families, clinicians and researchers.
If you are building an AAC app, the runtime saves you the hardest, least differentiating part. If you work with someone who uses AAC, the questions worth asking us are about languages, profiles and what runs without a connection. If you research communication, the behaviour and its limits are written down to be argued with.
ContinueBack to the overview