Semantic.art
multimodal hybrid-search vector-search semantic-routingSemantic.art
Semantic.art turns real, human-made art discovery into a multimodal search experience using
feelings, phrases, and images. It’s built with LanceDB hybrid search and semantic routing.
Learn how it's built
Read in detail about how Semantic.art is built in this blog post.
Wikipedia 41M Hybrid Search
multimodal hybrid-search vector-search ftsWikipedia 41M Hybrid Search
Interactive hybrid search, full-text search (FTS) and vector search demo with 41M+ Wikipedia entries.
Explore the power of combining FTS with vector search for more relevant results.
Learn how it's built
Read in detail about how the Wikipedia 41M hybrid search demo is built in this blog post.
Video Search
multimodal video-search vector-searchVideo Search
Search through a library of videos using natural language queries. This demo showcases how to use LanceDB
to perform semantic search on video content.
Large Scale Vector Search (1.5B)
vector-search education bge-embeddingsLarge Scale Vector Search (1.5B)
Semantic search over curated FineWeb EDU pages, represented as 1.5B+ BGE embeddings. Inspect the
query plan to see how LanceDB serves vector search at billion scale.
Robot Data Curation (DROID)
multimodal robotics vector-search ftsRobot Data Curation (DROID)
Search robot video frames from the DROID dataset, compute features, and select training examples in LanceDB.
Add captions and quality scores as columns, then write model errors back to guide the next training run.