AI module generation
Bedrock generates exercises based on aphasia type, severity and linguistic level. You stay in control via review & publish.
Early access · Spring 2026
AI-personalised exercises for every aphasia patient in minutes.
Aphasia Care generates clinically-grounded exercises tailored to your patient's aphasia type, severity and personal context — across twelve exercise types. You review, publish, and track progress.
eu-central-1·GDPR·MDR-aware roadmap
aphasia.care/modules/new
Module prompt
…
Exercise 1
Phoneme · /b/ in onset
Exercise 2
Word · name 'coffee'
Exercise 3
Sentence · 'I drink ___'
Exercise 4
Discourse · Morning routine
12
Exercise types
8
Aphasia types
NL/EN/DE
Content
EU
Hosting · GDPR
One platform, two perspectives
Build a complete exercise set in minutes, share patients with colleagues, and track progress from a single dashboard.
- Generate a module in ~2 minutes
- Review every exercise before it reaches the patient
- Track progress per patient and per module
Everything aphasia care needs to go digital
One workspace for generating, reviewing, practicing and monitoring.
12 exercise types
From naming and comprehension to spoken repetition, sentence ordering and functional dialogues — twelve types covering every modality.
Record & review speech
Patients record their speech and self-rate; you listen back, score and add notes. Recordings auto-expire after 180 days.
Cueing hierarchy
Built-in hints — semantic → phonemic → orthographic — support errorless learning, with a family mode that reveals the cue script.
Personalised to the patient
Exercises draw on the patient's occupation, hobbies, family and home language — patients submit their own context for your approval.
Collaboration
Multiple clinicians per patient with owner / collaborator / read-only roles and managed authority handover.
Progress analytics
Session-level scoring, time on task and answer patterns — aggregated per patient and per module.
Secure device binding
WebCrypto keys bind each patient device to an account. Lose a tablet, not access.
GDPR & MDR-aware
Hosted in eu-central-1, single-tenant data isolation per practice, audit trail on every handover.
Twelve exercise types, one generator
Every type adapts to the patient's aphasia profile, level and goals — generated, reviewed and published in one flow.
Naming
Name a pictured or described item — core word-retrieval practice.
Spoken repetition
Hear or read the target and say it aloud; the recording goes to the clinician.
Repetition
Repeat a spoken word or short phrase and confirm the match.
Comprehension
Pick the option that matches a sentence or question.
Sentence completion
Fill the blank in a sentence with the right word.
Word association
Choose the word that semantically fits the prompt.
Picture–word match
Match a word to the right image.
Auditory discrimination
Tell similar-sounding words apart.
Yes / No
Answer simple statements — low load for severe aphasia.
Sentence ordering
Arrange shuffled words into the correct sentence.
Functional script
Work through a real-life dialogue, turn by turn.
Semantic feature analysis
Explore a word's meaning, feature by feature.
Try a few exercises
This is the calm, large-button interface your patient sees. Pick an answer — it responds just like the real thing.
aphasia.care
Example 1 / 4
What is this?
☕
How it works
- Step 01
Patient profile
Capture aphasia type, severity and personal context (occupation, hobbies, home language).
1 - Step 02
AI generates
Bedrock picks the right mix of exercise types and composes a complete set — with cueing strategy, TTS audio and images.
2 - Step 03
Review & publish
Edit and publish. The patient practises and records on the bound device; missed items come back and results flow to you.
3
A module comes to life
A sneak peek of a personalised exercise set in our review screen.
ModulePublished
Morning routine · word level
Patient: Marie · Broca · moderate
- 01Make coffee
- 02Set the table
- 03Grab the newspaper
- 04Look outside
Cueing
Phoneme cue → semantic cue → modeling. Adapted to a Broca profile.
Generated in
42seconds
Estimated session
12min
Record at your own pace
Some exercises ask for real speech. The patient records, rates how it went, and the clinician listens back later.
Listen and repeat
apple
Record
How did it go?
Went well
Was difficult
Didn't work
Built in the EU, for clinical use
- eu-central-1 (Frankfurt) hosting
- Single-tenant DynamoDB isolation per practice
- PITR backups · Audit log on every change
- GDPR-aligned · MDR-aware roadmap
Join the early-access list
We're working with clinicians on the first release. Drop your email and you'll be first in line.
- No spam — only Aphasia Care updates
- Your data stays in the EU (eu-central-1)
- Early-access partners help shape the roadmap
Frequently asked questions
We're rolling out gradually in 2026 with a group of clinician partners. Add yourself to the early-access list.
All data lives in eu-central-1, single-tenant per practice, with audit trails on every change. We're GDPR-aligned and pursuing MDR-aware documentation.
Yes — every generated module passes through a review screen where you can edit text, audio and cueing before publishing.
A laptop for the clinician and a tablet or laptop for the patient. Devices bind securely via WebCrypto keys.
Yes — content can be generated in NL, EN and DE. More languages will follow as partners request them.
Twelve types — from naming, comprehension and sentence completion to spoken repetition the patient records themselves, sentence ordering and functional dialogues. The generator picks the right mix for the patient.
Recordings stay in eu-central-1, are only visible to the patient's care team, and auto-expire after 180 days.
Yes — a family mode reveals the correct answer and cueing script so a relative can guide the session without turning it into a test.
Pricing will be set ahead of general availability. Early-access partners get preferential rates.
Ready to accelerate your care?
Join clinicians shaping the future of aphasia therapy.