LanceDB
Embedded multimodal vector database on the Lance format
LanceDB is profiled here as a Vector Database tool for engineering teams. Read about features, pricing, and how it compares to related options in the tools directory.
Description
LanceDB is an open-source vector database built on the Lance columnar format, founded in 2022 by Chang She and Lei Xu. It runs in-process like an embedded database and stores data directly on object storage, which separates compute from storage and keeps large multimodal datasets cheap to hold. Versioned tables and zero-copy reads make it a fit for AI workloads that combine retrieval with training data management. The Lance format supports time-travel queries, so a table can roll back to an earlier version for reproducible experiments or audits.
Key Capabilities:
Embedded, serverless operation with no separate database server
Storage on S3, GCS, and Azure through the Lance columnar format
Vector, full-text, and hybrid search with metadata filtering
Multimodal storage for text, images, and embeddings together
Automatic data versioning with zero-copy reads
Apache 2.0 license with Python, TypeScript, and Rust APIs
Alternative tools
- Deep Lake
Database for AI that stores tensors and embeddings
- Marqo
Open-source vector search with built-in embedding inference
- Chalk
Feature platform for real-time machine learning data
- Hopsworks
Feature store and ML platform for batch and real-time data
- turbopuffer
Serverless vector and full-text search on object storage
- pgvector
Vector similarity search as a Postgres extension