Client ProjectCompleted5.0
From SaaS platforms to AI integration – here's a selection of my successfully completed projects.
Client projects and own developments
Let's build a modern web solution together.
Development of a SaaS training platform consisting of a trainer web app (Next.js), a client app (React Native / Expo) for iOS and Android, and an admin panel. Includes a Supabase backend, PDF export, push notifications, and submission to the App Store.
Mobile-first AI chat companion for food decisions. PWA with AI-powered chat that helps users make calm, clear food and movement decisions in real time — no calorie tracking, no dashboards, no guilt. Features: 6-question conversational onboarding, behavioral summaries, Stripe subscription ($19/month with 7-day free trial), push notifications, milestone system, admin panel with full retention dashboard. Tech: React PWA, Node.js, PostgreSQL, Stripe, AI provider abstraction (OpenAI/Anthropic).
Omni.Compare ist eine voll funktionsfähige, als Software-as-a-Service (SaaS) konzipierte Webanwendung, die es Benutzern ermöglicht, detaillierte und dynamische Vergleiche zwischen verschiedenen Produkten, Dienstleistungen oder Konzepten zu erstellen, zu teilen und zu entdecken. Die Plattform wurde mit einem modernen Tech-Stack entwickelt, der auf Next.js, TypeScript und Supabase für das Backend setzt. Ein zentrales Highlight des Projekts ist die Integration einer künstlichen Intelligenz, die den Erstellungsprozess revolutioniert. Die KI kann basierend auf den eingegebenen Elementen automatisch relevante Vergleichskriterien und passende Alternativprodukte vorschlagen.
PenSnap is the first dedicated AI-powered scanner for fountain pens and luxury writing instruments. Collectors, heirs, and thrift/estate-sale hunters take a single photo and receive an accurate identification within seconds — brand, model, variant, production era, nib material and size, and filling system — along with a market-value estimate based on real sales data. Unlike generic "identify anything" apps, PenSnap goes deep into one niche: it distinguishes a Montblanc 146 from a 149, a Pelikan M600 from an M800, and surfaces real marketplace sale prices instead of guessed values. A core architectural principle separates identification from valuation: the AI handles only identification (its strength), while prices come from a cached lookup keyed on model and condition, guaranteeing consistent, repeatable valuations rather than the fluctuating numbers that plague competitor apps. Built-in collection management, a wishlist with price alerts, and price tracking turn PenSnap from a one-time tool into a daily-use app. The backend follows a secure architecture where the AI API key never touches the client — every scan runs Client → Supabase Edge Function → AI API → back — enabling centralized rate-limiting, caching, and cost control at sub-cent per scan via a two-tier model strategy.
Eine moderne Cross-Platform-App (iOS/Android) zur Digitalisierung von Kitas. Bietet Echtzeit-Anwesenheitstracking, digitale Tagesrapporte, Event-Management und direkte Kommunikation zwischen Erziehern und Eltern. Entwickelt mit React Native, TypeScript und Supabase
