How I built an entire
app company — solo.

Most productivity apps follow the same playbook. Create an account, sync to the cloud, let the company monetize your data. Weight trackers selling health data. Budget apps analyzing spending for advertisers. Period trackers with privacy policies making headlines.

I kept looking for simple tools that just work without all of that. They didn’t exist. So I decided to build them myself. Not just one app, but an entire ecosystem. That decision is where Sapplify started.

Health data is personal. Weight trends, cycle patterns. Machine learning can add real value here by helping users understand their own data and presenting it in a way that actually makes sense. But the obvious path, sending that data to cloud models, would destroy the local-first approach. You lose control over where the data goes and who processes it.

So instead of using external ML frameworks or cloud APIs, I built custom statistical models in pure Dart. sWeight uses linear regression with confidence scoring and trend analysis. sCycle uses weighted prediction engines, anomaly detection, and pattern recognition for cycle and fertility forecasting. Everything runs on device, trained on the user’s own data, with zero network calls.

The apps are local-first, but the ecosystem still needs infrastructure. Instead of one monolith, I built two separate Python backends that share a Supabase database.

The public-facing backend runs on Flask. It handles the waitlist, beta access program, contact form, app feedback, analytics, and payment links. The second backend runs on FastAPI and powers sConsole. App Store Connect data sync, AI-powered content generation, image and video pipelines. Both use Supabase for auth and PostgreSQL, deployed separately so each can scale independently.

Marketing as a solo developer means automation or nothing. I initially used Make.com for content distribution, but it was too limited. So I built my own content pipeline directly into sConsole’s FastAPI backend.

The system uses Claude for text generation, Google Gemini for images, Veo 3 for short-form video, and OpenAI Whisper for captions. I create a topic, the backend generates a 14-day content calendar with platform-native formats. Carousel posts for Instagram, reels for TikTok, pins for Pinterest. All media gets uploaded to Bunny.net CDN automatically. What used to take hours of manual work now runs from one screen inside sConsole.