Back to Gap Map Infrastructure / Storage & Retrieval

LaminDB

Lamin Labs
open source / Overall score: 2.4

LaminDB is an open-source data management library for traceable, multimodal AI and life-sciences data. It queries, traces and governs datasets across diverse storage formats and locations through a Django-ORM registry, tracking lineage, branching and versioning over files, tables and arrays without a separate server.

lamindb on PyPI is the library itself - there is no separate server - and its metadata points back at laminlabs/lamindb, so it is the measurement identity and adoption reads from its downloads.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public
core features withheld
no

Apache-2.0 license body confirmed. The repository is public and unarchived and builds the whole library, and the README describes no paid tier, enterprise edition or license-gated build beside it - LaminHub is a separately sold hosted service rather than a piece withheld from the published library.

Adoption

2 high confidence
2.0

LaminDB's adoption reading comes from PyPI downloads of the `lamindb` package, since LaminDB is a Python library with no separate server, so the package is the product itself rather than a client to one. This is a direct signal of use, though it counts only PyPI installs and not use through other channels.

Capability

3 medium confidence
3.0

LaminDB is a metadata and lineage registry, backed by SQLite or Postgres, that tracks, versions and federates files, tables and arrays across pluggable storage locations rather than storing the data itself. That puts it alongside DVC: both provide lineage, versioning and governance over storage they do not themselves provide, rather than a distributed storage system in their own right.

  • https://github.com/laminlabs/lamindb/blob/main/README.md recorded 2026-10-07

    README: LaminDB "generalizes core lakehouse guarantees — ACID transactions, time travel, and schema evolution — to multimodal data", with metadata in SQLite/Postgres and data in parquet, zarr and other formats across local, S3 and GCP storage, and no storage engine of its own.

Verified 2026-09-24