AI Potluck
Infrastructure / Storage & Retrieval

LakeSoul

LakeSoul

LakeSoul is a cloud-native lakehouse table format and storage framework with a Rust-native metadata and file-IO core, ACID transactions, upsert semantics, schema evolution, automated compaction and fine-grained access control, shared identically across Spark, Flink, Presto, Ray, Daft and PyArrow bindings. Its Vortex file format can store multimodal data and vector embeddings alongside tabular data.

Verified 2026-09-02 via the GitHub API and the LICENSE body.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public
core-gated
ungated

Apache-2.0 license body confirmed at the repository root (the repository also carries REUSE-convention SPDX headers on individual files, but a root LICENSE exists and is the verbatim Apache-2.0 text). The repository is public and unarchived and builds the whole framework, and the README describes no paid tier, enterprise edition or license-gated build beside it.

Adoption

2 low confidence
2.0

3,248 GitHub stars, which lands in the 1K-10K stars band of the stars scale, level 2. LakeSoul ships as a table format consumed through Spark, Flink, Presto, Ray, Daft and PyArrow rather than a single installable package, and no deployment count is published, which leaves the stars scale and its cap of 3.

Capability

4 medium confidence
4.0

Banded on the category feature matrix as complete retrieval system, distributed or cluster-capable. Placed one band below the vespa anchor, alongside juicefs - LakeSoul owns its own metadata management, transaction log and compaction service rather than versioning storage it does not provide, which distinguishes it from lakefs and dvc at the band below. It supports vector retrieval as a file-format capability rather than hosting a ranking or embedding model in a serving path, so it does not reach the top rung.

Verified 2026-09-02