AI Potluck
Back to Gap Map Model components / Federated learning

FedJAX

Google
open source / Overall score: 1.5

JAX-based library for federated learning simulations, built for research ease of use, with federated datasets, models and algorithm implementations, and fast simulation on accelerators.

Not an officially supported Google product, per its README. The last release, 0.0.17, dates from 2023-07-12; the repository was pushed 2026-08-06.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI
source
public(the whole library)
core features withheld
no — a research library with no paid edition

FedJAX is an Apache-2.0 research library, and no paid tier exists for anything to be held back for.

Adoption

1 medium confidence
1.0

PyPI downloads of `fedjax`, the library itself.

Capability

2 high confidence
2.0

FedJAX simulates federations for research and offers no way to deploy one, one step below Plato, which can run over a real network.

Verified 2026-09-27