LanceDB
lancedb.com
- Free plan
- Linux
- API
- Self-hosted
What LanceDB is
LanceDB describes itself as an AI-native multimodal lakehouse for developing training datasets. It combines raw bytes, enriched features, metadata, and embeddings in searchable tables. Search options include vector or semantic search, full-text search, and hybrid search with SQL filters. Dataset curation tasks include deduplication and identifying edge cases for labeling. Users can write Python UDFs locally and run declarative feature pipelines at petabyte scale, according to the product page. Writes are automatically versioned, allowing users to branch, roll back, and tag experiments without duplicating data. The open-source software supports local or storage-path connections. LanceDB Enterprise includes Search, Exploratory Data Analysis, Feature Engineering, and Training, and can enforce platform-level index and table limits and safety checks on operations such as merge_insert. Lance Namespace integrates with Apache Hive Metastore, AWS Glue Catalog, and Apache Spark. The company describes distributing workloads through Ray and KubeRay on on-premises or cloud clusters. Enterprise setup requires contacting the team for a cluster, credentials, and endpoint details; no Enterprise price is listed. The Trust Center lists GDPR, HIPAA, and SOC 2 Type 2 under compliance.
Who it suits
LanceDB suits teams developing AI training datasets that need search, curation, and feature engineering across multimodal data. The Enterprise option is for organizations seeking a configured cluster and platform-level guardrails.
Good to know, and what to watch
- Supports vector, full-text, and hybrid search.
- Automatically versions writes without duplicating data.
- Dataset curation includes deduplication and edge-case identification.
- Open-source option supports local or storage-path connections.
- Integrates with Hive Metastore, AWS Glue Catalog, and Spark.
- Enterprise setup requires contacting the team.
- No Enterprise price is listed.
- Sparse vectors are not supported.
Plans
Every plan and priceLanceDB Enterprise
LanceDB OSS
Open source
The verdict
LanceDB brings dataset storage, search, curation, and versioning into one platform, with an open-source option and a separate Enterprise suite. Teams considering Enterprise should account for contacting the team to arrange a cluster and obtain credentials and endpoint details.
What the maker publishes
- Cloud deployments
- The company describes distributing workloads through Ray and KubeRay on clusters running on-premises or in the cloud. source
- Curation
- The product supports dataset curation tasks including deduplication and identifying edge cases for labeling. source
- Data types
- The platform brings raw bytes, enriched features, metadata, and embeddings together in searchable tables. source
- Enterprise access
- The Enterprise quickstart says users should contact the team to set up a cluster and obtain credentials and endpoint details. source
- Enterprise guardrails
- Enterprise can enforce platform-level index and table limits and safety checks on operations such as merge_insert. source
- Enterprise suite
- LanceDB Enterprise comprises Search, Exploratory Data Analysis, Feature Engineering, and Training. source
- Feature engineering
- Users can write Python UDFs locally and run declarative feature pipelines at petabyte scale, according to the product page. source
- Integrations
- Lance Namespace supports integrations including Apache Hive Metastore, AWS Glue Catalog, and Apache Spark. source
- Open-source languages
- The Enterprise quickstart offers installation paths for Python, TypeScript, and Rust. source
- Product
- LanceDB describes itself as an AI-native multimodal lakehouse for training dataset development. source
- Search
- It supports vector or semantic search, full-text search, and hybrid search with SQL filters. source
- Security and compliance
- The LanceDB Trust Center lists GDPR, HIPAA, and SOC 2 Type 2 under Compliance. source
- Versioning
- The product page says writes are automatically versioned and users can branch, roll back, and tag experiments without duplicating data. source
Checked against the maker's pages on 4 Oct 2026.