Project
AI Image Search
ActiveA live MongoDB tutorial project that turns a folder of images into searchable metadata in a React and Next.js app using Ollama, MongoDB Search, Vector Search, and hybrid search.
Instead of relying on filenames, a vision model understands each image, generates structured metadata, and stores everything in a single MongoDB document. The live demo at images.seemongo.com is a React and Next.js app that layers keyword search, semantic vector search, and hybrid retrieval with $rankFusion.
- Role
- Designer, builder, and tutorial author
- Technologies
- MongoDB Search, MongoDB Vector Search, Ollama, Voyage AI, React, Next.js
Highlights
- Live React and Next.js demo with keyword, vector, and hybrid search across AI-generated image metadata.
- Companion repo and how-it-is-built guide mapped to the official MongoDB for Developers tutorial.
- Single-document architecture: metadata, embeddings, and search indexes in one MongoDB collection.