Open-source lineage-native data lakehouse for traceable, multimodal AI with biological format support, built by the creators of Scanpy.
Lamin is an open data platform providing a lineage-native, format-agnostic lakehouse for tracing data, code, and agents across AI workflows. It supports querying and batch-loading across many datasets, manages metadata in registries and relational sheets, enforces schemas, and versions data and code like software with git-style change requests from agents and collaborators. Fine-grained permissions work at the database and storage level on AWS, GCP, or on-premises. Includes native support for biological data formats and registries. Lamin is a product of Lamin.
For people
For agents