{
  "slug": "lancedb",
  "name": "LanceDB",
  "url": "https://lancedb.com",
  "cat": "search",
  "group": "back",
  "tagline": "Embedded multimodal vector DB",
  "price": "OSS free (Apache 2.0), embedded, no server. LanceDB Cloud is usage-based across writes, queries and storage with $100 in free credits on the first month. Enterprise is custom.",
  "tier": "mixed",
  "pick": false,
  "what": "An embedded vector database built on the Lance columnar format — no server process, data sits as files on local disk or in S3. Does vector similarity, full-text search and SQL filtering over the same table, and stores raw bytes, images and video alongside embeddings. Python, TypeScript and Rust SDKs.",
  "why": "Think SQLite for vectors: <code>pip install lancedb</code>, point it at a directory, done. Zero-copy automatic versioning lets you time-travel an index, and the same code scales to the cloud tier.",
  "warn": "Embedded mode has no multi-writer coordination — one writer at a time, or move to the managed tier. Cloud per-unit rates are behind a calculator rather than published.",
  "install": "pip install lancedb",
  "category": {
    "key": "search",
    "name": "Search, vector & RAG",
    "desc": "Keyword search, vector stores, and web-search APIs for agents — three different things.",
    "group": "back",
    "groupName": "Give it a backend"
  },
  "freshness": null,
  "health": {
    "ok": true,
    "status_code": 200,
    "final_url": "https://www.lancedb.com/",
    "checked_at": "2026-08-18T17:12:38.21+00:00"
  },
  "corrected_at": null,
  "upstream_changed_at": null
}