
AI-Native App: FemCycle
Solo-engineered end-to-end design, writing, and management
FemCycle is a health app developed from concept to deployment using AI - by me. Built in React Native with Cursor, Claude, and Gemini, the project includes the core mobile experience, a content design system, and custom management tools. Development is ongoing.
-
Phase 3: Conversational Q&A flow (coming soon)
-
Phase 4: User testing (coming soon)
-
Phase 5: UX writing agent (coming soon)
Background

Why
Cycle trackers allow women to document their cycles to get to know their body and manage health and fertility. However, existing apps are costly and contain more features than many users need.

What
A clean, lightweight, customizable cycle tracking app for women who need basic cycle tracking features. The app is designed, written, and built by me using AI.

How
I worked with Gemini to strategize and used Cursor and Claude to build and create a content system and management tools for a future-ready infrastructure.
Phase 1: MVP
Product & content strategy
Initial user-needs exploration for MVP

Initial product, UX, and UXW definition

Work with Gemini to build implementation plan

Iterative work with Cursor & Claude to create the MVP
Process
Tools: Gemini, Cursor, Claude Code
Main MVP screens
In this phase, the UI copy was fine-tuned by me while error and info dialogs were written by Cursor.
Data input screen

Data input screen (Cont'd.)

Visual chart

Calendar view

Phase 2: Content design system
Content guidelines

Content audit

Implementation

Ditto integration

Documentation on Notion
Tools: Cursor, Claude Code, Ditto, Notion
Process
Content guidelines
I used Claude to automatically catalog the copy in the codebase and create a voice and tone guide from scratch, which I then fine-tuned manually.
Content audit
I used Claude to review the app copy against the guidelines and flag inconsistencies, potential issues, and missing messaging. I reviewed the audit, made decisions and corrected as needed.
Some flagged terms are to be tested with users before a final decision.

Audit implementation
I uploaded the final audit to Cursor for fixing, and manually reviewed the changes to ensure correct implementation.
Examples:
Before



After



Ditto integration
I connected the codebase to Ditto (a content management tool) to allow me to manage the strings independently without needing help from Eng.

Documentation on Notion
I uploaded the content artifacts and decisions to Notion for future reference and usage by other writers.

Next steps?
-
Phase 3: Conversational Q&A flow (coming soon)
-
Phase 4: User testing (coming soon)
-
Phase 5: UX writing agent (coming soon)
