# LanceDB

> AI-native multimodal lakehouse for vector search, data storage, and model training at petabyte scale.

LanceDB is an open-source AI-native multimodal lakehouse designed to handle complex AI workloads from agents to models, search to training. It provides a unified platform for storing, searching, and processing multimodal data at petabyte scale, eliminating the need for separate data lakes, vector databases, and search engines.

The platform combines storage, search, feature engineering, analytics, and training capabilities into one cohesive solution, enabling teams to move quickly from prototype to production.

- **Multimodal Data Storage** uses a new columnar standard optimized for multimodal data with fast scans, random access, large blob storage, and zero-copy fine-grained data evolution at petabyte scale.

- **Advanced Hybrid Search** delivers blazing fast hybrid search, filtering, and reranking over billions of vectors with compute-storage separation for up to 100x cost savings.

- **Automated Feature Engineering** provides declarative, distributed, and versioned pre-processing for faster feature experimentation with native support for LLM-as-UDF.

- **High Performance SQL** enables exploration, curation, and analysis of multimodal data with intuitive SQL queries.

- **Optimized Training Pipelines** offer faster dataloading, global shuffling, and integrated filters for large-scale training using PyTorch or JAX.

- **Serverless Cloud Option** provides a managed serverless retrieval engine with automatic indexing, compaction, and an intuitive UI to explore data.

- **Enterprise-Grade Compliance** includes SOC2 Type II, GDPR, and HIPAA compliance for data safety and security.

- **Multi-Language SDKs** support Python, TypeScript, and Rust for flexible integration into existing workflows.

To get started, sign up for LanceDB Cloud to access the serverless search engine with $100 in free credits. Connect to LanceDB, ingest your data, and build indexes to streamline your workflow. For enterprise deployments with complete data control and multimodal SQL engine capabilities, contact sales for custom pricing and deployment options on any cloud.

## Features
- Multimodal data storage
- Hybrid vector search
- Automatic indexing and compaction
- Feature engineering pipelines
- High performance SQL for multimodal data
- Optimized training pipelines
- PyTorch and JAX integration
- Zero-copy data evolution
- Compute-storage separation
- Distributed data pre-processing
- LLM-as-UDF support
- Global shuffling for training
- Petabyte scale support
- Data lake compatibility

## Integrations
PyTorch, JAX, Hugging Face, Python, TypeScript, Rust, AWS S3

## Platforms
WEB, API, DEVELOPER_SDK

## Pricing
Freemium — Free tier available with paid upgrades

## Links
- Website: https://lancedb.com
- Documentation: https://docs.lancedb.com/
- EveryDev.ai: https://www.everydev.ai/tools/lancedb