For clinicians
Ryan's Voice for speech pathologists
Everything you need to decide whether this belongs on a client's shortlist — including where it doesn't. Written for clinicians, not marketing.
What it is: a phrase-level, conversation-aware AAC app for iPhone and iPad. Whole utterances on large tiles rather than word-by-word construction, with optional on-device AI that surfaces phrases matching the current conversational topic. Core app is free, so trialling costs nothing and needs no funding decision.
Feature matching, honestly
We would rather you chose the right system than ours. Here is where Ryan's Voice fits, and where it doesn't.
Consider it when the client…
- Is literate or emerging-literate, and finds symbol-first systems limiting
- Has motor-planning difficulty (apraxia) that makes dense grids error-prone
- Is a teen or adult who finds typical AAC vocabulary childlike
- Needs speed in real conversation more than infinite generativity
- Has acquired communication loss — stroke/aphasia, MND/ALS, brain injury
- Is in a setting requiring strict privacy or reliable offline function
- Has no funding yet and needs something today
Look elsewhere when the client…
- Needs switch scanning, head pointing or eye gaze — we are touch-only
- Is an emerging communicator learning single words and symbol relations
- Is established on a taught motor-planning vocabulary such as WordPower or LAMP
- Needs deep board and grammar customisation
- Requires Android, Windows or a dedicated device
- Needs robust, fully generative morphology and syntax construction
Vocabulary provenance
The library is 1,120+ phrases across 27 conversation topics, sourced from years of real speech-therapy sessions with one literate, minimally-speaking teenager — not written by copywriters. It includes utterances such as "She seems disappointed by the outcome," at genuine conversational and emotional depth.
An internal audit found 100% coverage of the Universal Core 36 (Center for Literacy and Disability Studies, UNC Chapel Hill) and of the Beukelman, Jones & Rowan (1989) preschool frequency list. Two of the 74 words in the combined set are covered by Australian-English equivalents rather than the US form, so the raw figure is 97.3% and 100% dialect-adjusted. That audit is an internal document, not peer-reviewed, and we would rather say so than imply otherwise.
Built for clinical review
- Optional session log. When enabled, the app records which phrases were used during a flagged therapy session so you and the client can review them together. It stays on the device, and the client, parent or carer can view, clear or disable it at any time in Settings → Privacy.
- Provenance is visible per phrase. Users and clinicians can see whether an utterance came from the curated library, was typed by the user, or was generated on-device — useful when you need to know what the person composed themselves.
- Nothing leaves the device. No account, no cloud, no analytics. Microphone audio is processed in real time for recognition and discarded; it is never recorded or transmitted. This matters in schools, hospitals and any setting with data-governance requirements.
Access and design
Phrase tiles never drop below a 60pt touch floor, with primary tiles at 90–140pt — well above the 44pt platform minimum — and category changes use a horizontal swipe, a continuous motion many clients retain when precise pointing is difficult. Layouts show a small number of large tiles rather than a dense array. A symbol mode pairs the open-licensed Mulberry symbol set with text.
Current limitation worth stating plainly: there is no switch, scanning or alternative-access support today, and the app is iPhone/iPad only, iOS 18 or later.
Free clinician access
The core app is free to download and trial now. If you would like Smart Voice — the on-device conversation engine — unlocked at no cost so you can evaluate it properly with clients, tell us below and we will send access codes.
Download the free appRequest access, or tell us what's missing
We are actively looking for clinical feedback — particularly on symbol mode, which is deliberately basic and which we want to improve with SLP input. Blunt criticism is more useful to us than encouragement.