ProjectsB4
Anuah
A wellbeing app for couples, built on cycle context, a personal check-in and practical guidance.
- Role
- Creator / Product Lead — AI-assisted development
- Since
- April 2026
- Status
- In development, in internal testing on Android.
SummaryAerial view
Anuah helps a couple understand how one of them is doing today. It combines menstrual cycle context with a personal check-in and turns both into practical guidance for the partner. It started in April 2026 from a personal need.
A check-in changes the guidance
Synthetic dataTwo fictional screens. Change the check-in and read how the partner's briefing follows.
This demo needs JavaScript. The case below reads the same without it.
CaseOn foot
Problem
Knowing the phase of a cycle does not tell you how a person feels today, or how to be there for them.
Decisions
The check-in leads, the phase is context
BusinessWhat a person reports today about mood, energy and social battery counts more than any assumption about a phase. The app never replaces a real check-in with a stereotype.
TechnicalThe daily briefing is generated from the check-in plus the cycle context. The check-in is a required input, not a decoration.
Two connected experiences
BusinessOne person gets self-knowledge. The linked partner gets orientation they can act on: how to talk, what to plan, a small gesture.
TechnicalAn Expo and React Native app on Supabase, with linked accounts and explicit rules about what each side can see.
Context is built on the server
BusinessWhat reaches the AI is decided by rules, not by whatever the app happens to send.
TechnicalSupabase Edge Functions assemble the context and call the Claude API. The client never talks to the model directly.
How I checked it
An Android build is in internal testing and in private use.
What I learned
In a sensitive product, what you decide not to claim is part of the design. The limits came first, and the features had to fit within them.
Role
Concept, journeys, copy, decisions about context and permissions, priorities and acceptance criteria. The implementation was assisted mainly by Claude Code.
EvidenceBelow the surface
Architecture
- Person
- Data
- Storage
- Service
- AI model
The same flow, step by step
- One person completes a check-in. Cycle context is stored alongside it.
- Supabase keeps both, with rules about who can see what.
- An Edge Function builds the context and calls the Claude API.
- The briefing reaches the linked partner as practical guidance.
Product plates
Plate 1 of 6Opening screenEditorial compositionSynthetic data Plate 2 of 6Cycle contextEditorial compositionSynthetic data Plate 3 of 6Understanding the cycleEditorial compositionSynthetic data Plate 4 of 6Partner briefingEditorial compositionSynthetic data Plate 5 of 6Check-inEditorial compositionSynthetic data Plate 6 of 6The connected experienceEditorial compositionSynthetic data
Stack
Page last revised: