AI Potluck
Contribute
Back to Gap Map Infrastructure / Storage & Retrieval

Feast

Feast project
open source / Overall score: 3.0

Feast is a feature store that keeps one definition of a feature and serves it from two places: an offline store for building training sets with point-in-time correctness, and an online store for low-latency retrieval at inference. Materialization moves values between them, and a registry holds the definitions.

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 product, and the README describes no paid tier, enterprise edition or license-gated build beside it, so source is public and the core ungated.

Adoption

3 high confidence
3.0

PyPI downloads of feast drive this reading, since that package is the channel Feast declares for distribution. The figure therefore measures use of the tool itself, not a proxy for it.

Capability

3 medium confidence
3.0

Feast orchestrates point-in-time-correct offline retrieval, online serving and materialization across pluggable stores, with a feature registry. Its alpha vector search calls into a separate vector store rather than providing an index of its own, which keeps it two steps below Vespa, a complete distributed retrieval system.

Verified 2026-09-24