LanceDB
LanceDB stores multimodal data in versioned tables for data curation, feature engineering, search, retrieval, and model-training workflows.
What it does
- Vector search
- Knowledge retrieval
Features 5
- Schema-enforced multimodal tableKeeps episodes, frames, metadata, features and embeddings together in one schema-enforced table.
- Vector, full-text and hybrid searchRuns vector, full-text, hybrid and filtered search, reranking and SQL queries on a table.
- Data curation operationsProfiles, samples, chunks, joins and deduplicates records before they are used for training.
- Computed columns and materialized viewsDerives columns and views from existing data and refreshes them as the data changes.
- StreamingDataset for PyTorchStreams multimodal training data into PyTorch as globally shuffled, deterministic batches.
Sources
- Pricing
-
Pricing page: "Free and open source — pip install lancedb"
- Description
-
Official documentation · 5 Nov 2026
Product page: "one data layer for curation, feature engineering, search and retrieval, and model training."
- What it does
-
Official documentation · 5 Nov 2026
Product page: "one data layer for curation, feature engineering, search and retrieval, and model training."
- Sold within
-
Sold on its own with published plans.
- Name
-
Official documentation · 5 Nov 2026
Name as published on the product page.
- URL
-
Official documentation · 5 Nov 2026
Product page, read 2026-11-05.
- Deployment
-
Official documentation · 5 Nov 2026
Product page: "Run LanceDB as a managed deployment or install it inside your own cloud account with BYOC."
- Interfaces
-
Official documentation · 5 Nov 2026
Product page: "Run LanceDB as a managed deployment or install it inside your own cloud account with BYOC."
Something wrong here? Send a correction — quote this product id: lancedb-platform.