Now testing on AndroidBeta

Learning that adapts to the child — not the other way around.

FlexiKeys adaptive keyboard screen: a spelling exercise showing a melon image and letter tiles, from the real app

Helping every child learn at their own pace. FlexiKeys teaches typing, literacy, and digital interaction to children ages 3–10 — built first for children with motor difficulties, and just as great for every child.

Parents · Rehabilitation centers · Inclusive schools · Speech & occupational therapists

Try BetaBeta

Android Beta · Free · Currently available for testing

Learn More

Tested with real children

Built with therapists' feedback

Adaptive learning, invisible to the child

MVP Built
First Field Validation Completed
Iterating with Real User Feedback

The problem

Most digital interfaces assume a level of motor precision a lot of children don't have.

Small buttons, fast timing, and single-input gestures aren't neutral design choices — they quietly decide who gets to use a product at all.

Cerebral palsy is the most common motor disability of childhood, affecting movement, balance, and posture.

CDC

It's a spectrum — many children also have differences in speech, vision, hearing, or attention alongside motor ability.

Damiano et al., 2021 · PMC9619294

Touch input can be affected by reduced precision, tremor, slower movement, and faster fatigue than standard interfaces assume.

FlexiKeys domain research

Guidance for assistive technology aimed at cognitive and communication needs — exactly this age group — remains thin.

Zhang & Borg, 2025 · doi:10.3389/fresc.2025.1581104

Why existing tools fall short

Access tools, communication tools, and learning tools rarely meet in one place.

In our own review of the space, we didn't find another product combining a self-adjusting motor-accessible interface with a real literacy curriculum for this age group. That's a specific gap, not a claim that no one else is working in assistive technology.

CategoryWhat it givesWhat it misses

Adaptive input hardware

e.g. specialized switches, adapted keyboards

Gives:Physical access to a device

Misses:No literacy curriculum, no learning progression

AAC communication apps

e.g. symbol- and speech-based communication tools

Gives:A path to communication

Misses:Not built to teach typing or literacy as a skill

Mainstream EdTech apps

e.g. general early-literacy apps

Gives:A learning curriculum

Misses:Fixed interface, one size fits all, no motor adaptation

→

FlexiKeys sits across all three: a motor-adaptive interface, wrapped around a real literacy curriculum, for children ages 3–10.

Introducing FlexiKeys

One platform, three plain answers.

What is it

An adaptive platform teaching typing, literacy, and digital interaction to children ages 3–10 — 16 learning levels plus a drawing and fine-motor module, in English.

Who it's for

Built first for children with motor difficulties like cerebral palsy — and, because the adaptation is invisible, equally good for any child learning at their own pace.

How it helps

The keyboard, pacing, and hints adjust automatically to each child's real performance — no manual configuration, no separate "mode" to switch on.

Charos Sherbekova, FlexiKeys founder, presenting to families and staff at a rehabilitation center

Why we exist

We built this because the gap in section three isn't abstract — it shows up in real classrooms and therapy rooms.

FlexiKeys is led by Sherbekova Charos (founder) and Abdulloh Ochilov (co-founder), with Madxiddinov Shoxjahon building the product. The team set out to answer one specific question: what would a literacy app look like if it were designed for a child with limited motor precision first, instead of retrofitted for them later?

That question shaped every decision in the product — from how the adaptive engine measures hesitation and touch offset, to the rule that a struggling child never sees a red X or a buzzer. It also shaped how we chose to validate the idea: by bringing an early version directly to a rehabilitation center and watching real children use it, rather than waiting for a polished version to test.

How it works

An engine that watches how a child actually performs, and adjusts within limits.

Every tap is measured — target vs. actual key, hesitation before the first touch, how far off-center a touch lands, and whether accuracy drifts down over a session. The engine turns those signals into small, gradual interface changes. Nothing jumps; a child never notices a sudden shift.

Step 1

Open FlexiKeys

Step 2

Complete adaptive activities

Step 3

System adjusts difficulty

Step 4

Track progress

What the engine seesWhat it adjusts

Accuracy stays low across a session

Key size grows gradually (up to 2×)

Response time runs high

Press-and-hold time increases, filtering slips

Accidental taps are frequent

Repeat-tap sensitivity is reduced

Accuracy is consistently strong

No change — adaptation is invisible until needed

In simulated testing of a struggling-child profile (50 keystrokes, ~38% accuracy), the engine grew key size to 1.35× and raised guidance level within the session — then left a fast-learning profile completely unchanged. That behavior is unit-tested, not just described.

FlexiKeys home screen showing the child's levels and mascot

Home & levels

Parent dashboard

Parent assistant

On the parent assistant: it answers questions about a specific child's real progress — accuracy, time practiced, what changed and why — grounded in that child's actual data. It gives educational guidance only and always discloses that it isn't a substitute for medical or therapeutic advice.

Product principles

What we hold ourselves to, on the record.

Accessibility first

Touch targets, timing, and error handling are designed for the hardest case first, then verified to work for every child.

WCAG 2.1 AA reviewed

Evidence driven

Every adaptation rule ships with a test proving the metric-to-adjustment mapping works, checked against real testing before we call it done.

Adaptive engine, bounded & unit-tested

Built with families

The product is shaped in the room with the children and caregivers who use it — testing sessions inform the roadmap, not just validate it afterward.

Field-tested with real families

Privacy by design

Children are identified by a pseudonymous ID, never a name or birthdate. No third-party analytics touch the child experience.

COPPA / GDPR-K aware

Evidence

What our first real-world test told us

We ran a structured usability session with children, parents, and caregivers at a rehabilitation center — observing, with a written protocol, how children used the adaptive keyboard and drawing activities largely on their own. This is one early test, not a clinical trial, and we treat it that way below.

A teenager using a tablet during a structured FlexiKeys usability session, with a mobility walker beside him and a printed Child Testing Observation Form on the table

A real session, not a staged photo: an observation form on the table, testing a child using his own mobility aid.

What we validated

Children were willing to engage with the application, and parents showed genuine interest in the concept. Even as an early MVP, the testing confirmed that the problem is real and that there is demand for a digital tool designed for children with developmental and learning differences.

What we learned

Children interacted with the product in different ways and at different speeds — some quickly understood the activities, while others benefited from additional guidance. We also learned that real rehabilitation environments are naturally dynamic, which makes clear onboarding and well-defined facilitator roles just as important as the product itself. The first session wasn't frictionless, and we're not presenting it as if it were — that honesty is the point.

A child closely supported by a caregiver while using FlexiKeys on a phone — One of the caregiver-supported sessions behind the "willing to engage" finding above.

Children and facilitators working through FlexiKeys drawing activities together at the rehabilitation center — The group drawing activity where pacing differences and facilitator roles became clear.

What we're improving

Following the first validation session, we are:

Simplifying the onboarding experience
Improving the first-time user flow
Refining learning activities based on direct observation
Improving our own user-testing process
Scheduling regular validation sessions every one to two weeks with children, parents, caregivers, and rehabilitation centers

Built with real families. Improved through real feedback.

Built to be trusted

We reviewed our own work against real standards, and published what we found.

For rehabilitation centers, schools, and reviewers who need to check this before bringing it to families — the summary is below, and the full detail is one click away.

Accessibility

Reviewed

Reviewed against WCAG 2.1 AA — including open items we haven't finished yet, listed honestly.

Security

Reviewed

Assessed against OWASP ASVS Level 2 across authentication, session handling, and access control.

Child privacy

Documented

COPPA / GDPR-K–aware data handling: pseudonymous child IDs, no third-party analytics in the child experience.

Read the full accessibility, security, and privacy detail

Who this is for

Find your own way in.

Parents

A companion that adapts to your child, quietly — never a separate mode, never a label your child can see.

Rehabilitation centers

Invite us to test with your families — we're scheduling sessions every one to two weeks and want your observations shaping the roadmap.

Inclusive schools

A curriculum built for a wide range of ability from the start, not adapted after the fact.

Therapists

Progress reporting in plain language, grounded in real data — built to support your work, not replace your judgment.

Partners & investors

See where the product is today, what we learned from real testing, and where it's going next.

Future team members

A small team building with real engineering discipline on a problem that matters. We're just getting started.

Vision

What's built, what's next, and what we're building toward — kept separate on purpose.

Today

16 learning levels — letters through stories — plus a drawing and fine-motor module
Adaptive engine: key size, timing, and hint level adjust to real performance
Parent dashboard and assistant, grounded in each child's own data
Built in English, Uzbek, and Russian

A simpler first-time onboarding flow, direct from our first testing round
Regular validation sessions every one to two weeks with families and centers
Expanding testing partnerships beyond our first rehabilitation center

Long-term

Every child, regardless of ability, reaching real digital independence
An adaptive engine that keeps learning alongside the child, not just the curriculum
A tool trusted by rehabilitation centers and schools as standard practice, not a novelty

Team