Appster: AI app discovery with a full stack content platform
Natural-language discovery across a catalogue of more than 2,700 apps, with semantic vector search, administration and programmatic SEO pages.
The product problem
People searching for an application do not always know its name or category. Appster lets them describe what they need in natural language. That experience requires a searchable catalogue behind the interface, alongside a way to administer the app data and expose useful catalogue destinations on the web.
The delivered scope
Appster indexes more than 2,700 apps with semantic vector search over an OpenAI-embedded catalogue. It includes a full admin panel and thousands of prerendered, programmatically generated SEO pages. Next.js, Supabase, TypeScript and Tailwind CSS form the application stack.
AI as one part of the product
The search capability sits inside a wider application: catalogue data, administration and public content pages all matter. A similar integration should be scoped around retrieval quality, content maintenance and the user journey, with acceptance criteria for the actual application rather than a model demo alone.
What shipped
A discovery platform combining semantic search across more than 2,700 apps, catalogue administration and programmatic content delivery.
Visit the project ↗