lancedb/lancedb

Developer-friendly OSS embedded retrieval library for multimodal AI. Search More; Manage Less.

What it solves

LanceDB is a multimodal data platform designed to handle the storage, indexing, and searching of massive amounts of multimodal data (such as text, images, videos, and point clouds) and their corresponding vectors. It provides a scalable, production-ready solution for AI/ML applications that need to perform fast vector similarity searches across petabytes of data.

How it works

Built on top of the Lance columnar format, LanceDB allows developers to store and query data using a combination of vector similarity search, full-text search, and SQL. It supports GPU-accelerated index building to speed up the process. The system is designed for zero-copy reads and automatic versioning, allowing users to manage data versions without additional infrastructure.

Who it’s for

AI/ML developers and researchers who need a high-performance vector database that can store both vectors and metadata, and integrates seamlessly with tools like LangChain, LlamaIndex, Pandas, and Polars.

Highlights

  • Multimodal Support: Handles vectors, metadata, and various data types like images and videos.
  • Fast Vector Search: Capable of searching billions of vectors in milliseconds.
  • Comprehensive Search: Combines vector, full-text, and SQL search capabilities.
  • Flexible Deployment: Available as an open-source local version or a managed cloud/enterprise service.
  • Broad Integration: Native SDKs for Python, Node.js, and Rust, and a REST API.

Related

  • Project
    lance-format/lance

    An open lakehouse format for multimodal AI that provides high-performance vector search, random access, and native storage for images, video, and audio.

  • Project
    ModernRelay/omnigraph

    Omnigraph is an open‑source graph database built on the Lance columnar format that runs on any S3‑compatible object store. It offers typed schemas, multimodal retrieval (graph + vector + full‑text), Git‑style branching, and Cedar‑based security policies, all managed via Terraform‑like `cluster` files. Designed for long‑horizon AI agents, it lets multiple agents share a versioned world model and retrieve context in a single query.

  • Project
    dingodb/dingo

    A distributed multi-modal vector database that integrates real-time strong consistency, relational semantics, and vector semantics into a unified platform.

  • Project
    infiniflow/infinity

    An AI-native database providing high-speed hybrid search across dense and sparse embeddings, tensors, and full-text for LLM and RAG applications.

  • Project
    ArcadeData/arcadedb

    A high-performance multi-model DBMS that supports graph, document, key-value, vector, and time-series data in a single ACID-compliant engine.