Skip to content
All work
04Cognitive Training · Web & Mobile

Cogniware

Turning assessment results into training people actually come back to.

Platform
Web + mobile
Stack
Next.js · TypeScript · React Native · Expo · Supabase · Stripe · Gemini API
My role
Full-stack product engineering: the web app and API, the mobile app, exercise runtime and scoring, personalization, payments, and native eye-tracking research.
usecogniware.com
Cogniware homepage: train your brain like you train your body, with today's plan and the four domain scores
01

Overview

Cogniware turns four cognitive domains — attention, timing, impulsivity and hyperactivity — into twelve short, adaptive exercises, from Rare Target Watch to Go / No-Go and Stop Signal.

One Next.js app serves the website, the dashboard and the API that the mobile app uses.

02

The problem

Assessment results like a Moxo report describe where someone struggles, but they rarely turn into a routine. Generic brain games, on the other hand, ignore those results entirely.

03

The approach

Make the assessment optional but useful. Anyone can start training immediately; people who enter their scores get a plan that balances the four domains and shifts emphasis as their performance changes.

Keep sessions short — as little as ninety seconds — and make progress visible: accuracy, reaction time and consistency after every run.

04

Engineering

Scoring and enrollment live in shared modules, so the web dashboard and the mobile app calculate results the same way, against one API and one database.

  • E.01

    Adaptive exercise runtime

    Twelve exercises share one runtime that adjusts difficulty to performance and records accuracy, reaction time and consistency for every session.

  • E.02

    From scores to a plan

    Moxo index scores can be entered by hand or extracted from a report with the Gemini API, then turned into a weekly plan prioritized around the four domains.

  • E.03

    Private by default

    Assessment reports are kept in private storage with tested access policies. There is no public bucket for personal documents.

  • E.04

    Eye-tracking lab

    A separate feasibility app tests gaze tracking with a Swift module on ARKit and a CameraX/MediaPipe module on Android, using nine calibration and five validation targets. It stays out of scoring until the data proves reliable.

  • E.05

    Subscriptions and email

    Stripe handles the trial and subscriptions, and Resend sends transactional email.

05

My role

I worked across the web app and API, the database, the mobile app, the exercise runtime, payments, and the eye-tracking research prototype.

06

Outcome

A live cognitive training product at usecogniware.com with twelve adaptive exercises, optional assessment-based plans, and a mobile app on the same API.

Cogniware's four training domains — attention, timing, impulsivity and hyperactivity — with three exercises each