Hello, the demo searches 10,000 paintings using semantic search, so you can type things like "dark forest with fog" or "people dancing at a party" and it actually finds relevant results even though none of those words appear in the metadata
The underlying product is Vecstore, a search api for text and images. the idea is you skip the whole embedding pipeline + vector database setup and just call an api. You send in your data, you search it. that's it
It also does reverse image search, face search, OCR (finding images by text inside them), and nsfw detection. all from the same api
kencho•1h ago
The underlying product is Vecstore, a search api for text and images. the idea is you skip the whole embedding pipeline + vector database setup and just call an api. You send in your data, you search it. that's it
It also does reverse image search, face search, OCR (finding images by text inside them), and nsfw detection. all from the same api
Built with Rust