Offline
100% on-device. No account, no cloud.
Every child
has a voice.
Mela turns taps into speech — a warm, simple AAC board for non-speaking children and adults.
Free. Forever. Non-profit.
eat
more
happy
play
love
help
Accessible
Switch scanning, big targets, calm motion.
Community-built
Shaped with families & therapists.
Who it's for
One app, three lives changed
Mela adapts to the person using it — from a 5-year-old autistic child pointing at pictures, to a parent with ALS typing a last story for their kids, to a grandfather relearning words after a stroke.
- Children with non-verbal autism — symbol grids with core vocabulary, motor-planning friendly
- People with ALS — predictive keyboard, voice banking, progressive input fallback (switches, eye tracking)
- Stroke survivors with aphasia — dual symbol + text mode, forgiving phrase recall
- Anyone with speech challenges — apraxia, cerebral palsy, post-laryngectomy, selective mutism
A voice they recognize
Their voice, in a familiar tone
A parent, a partner, a caregiver can record a short sample of their own voice — and lend it to the person who needs it. Mela synthesizes every phrase on-device in a voice that feels like home. No cloud. No subscription. No data collection.
- Clone from as little as 30 seconds of speech
- 100% on-device synthesis — never uploaded
- Explicit, revocable consent from every voice donor
- Tiered quality: Basic (30s) to Perfect (3+ min)
- Multiple modes per donor: warm, firm, gentle, urgent
Their words, their world
Speaks like your family does
Mela doesn't just borrow a voice — it borrows the way your family speaks. Nicknames, pet names, the specific words you use for food or sleep or love. On supported devices, on-device AI rephrases every sentence in your family's style, approved by you before it's ever used.
- Custom family lexicon: "Mum" not "Mom", "cuddles" not "hugs"
- Style-matched phrasing powered by Apple Intelligence, on-device
- Every rephrased phrase reviewed and approved by the parent
- Dual mode: "family" voice vs "everyone" voice for outsiders
- Export family glossary for school, therapists, and caregivers
Thoughtful throughout
Built on AAC best practices
Mela isn't a bright new idea; it's every well-established AAC principle, made free, beautiful, and modern.
Symbol grid
ARASAAC and Mulberry symbol sets — thousands of pictograms, free to use. Motor planning-friendly layouts from 4×4 to 10×10.
Predictive keyboard
For literate users: fast typing, phrase prediction, and sentence scaffolding that respects your preferred vocabulary.
Eye tracking ready
Works seamlessly with iOS 18 Eye Tracking, Switch Control, Voice Control, and AssistiveTouch — progressive for ALS users.
Bilingual from day one
English and French, with the same care in both. Regional variants (UK, US, Australian, Quebecois) respected.
Works offline
Once the voice is banked, Mela works entirely offline. School buses, airplanes, long car rides — no signal required.
No data ever leaves
No cloud, no analytics, no tracking, no ads. Voice recordings, models, vocabulary — all stay on the device. Forever.
Why free
Communication is not a luxury
The current AAC apps cost $250 to $400. Families already stretched by the cost of disability are asked to pay again for the right to be heard. We think that's wrong. Mela is free because the families who need it most are the ones who can't afford to pay.
LorisLabs is a solo, self-funded studio. Mela will never have ads, never sell data, never require a subscription. If we need to cover costs, we'll do it via transparent donation — never by gating communication.
