pfl-research
AppleApple's Python framework for private federated learning simulations. Researchers run fast simulations across processes, GPUs and machines, with local and central differential privacy mechanisms built in and benchmark suites over realistic datasets.
The PyPI package is `pfl`; its project URLs name apple/pfl-research. Apple states that the framework is not intended for third-party federated learning deployments.
Openness
5 high confidence- license
- Apache-2.0(OSI
- source
- public(the whole framework)
- core features withheld
- no — a research framework with no paid edition
pfl-research is Apache-2.0 and exists for research simulations, so there is no paid tier or commercial deployment product to hold anything back for.
- https://raw.githubusercontent.com/apple/pfl-research/HEAD/LICENSE recorded 2026-09-27
The LICENSE body: the verbatim Apache License, Version 2.0 text with no added clause.
- https://raw.githubusercontent.com/apple/pfl-research/HEAD/README.md recorded 2026-09-27
README: "The framework is `not` intended to be used for third-party FL deployments". No paid tier is named.
- https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/apple%2Fpfl-research recorded 2026-09-27
Repository record for apple/pfl-research: license apache-2.0, not archived, not a fork, pushed 2026-09-16, 358 stars.
- https://ungh.cc/repos/apple/pfl-research/files/develop recorded 2026-09-27
The 439-file tree of apple/pfl-research at `develop`: one root LICENSE, and no ee/, enterprise/, commercial/, proprietary/ or premium/ path anywhere in it. Path absence only, paired with the statement cited beside it.
Adoption
1 medium confidencePyPI downloads of `pfl`, the framework's package.
- https://pypi.org/pypi/pfl/json recorded 2026-09-27
PyPI `pfl` 0.5.2, uploaded 2026-09-16; homepage and repository https://github.com/apple/pfl-research.
- https://pypistats.org/api/packages/pfl/recent recorded 2026-09-27
last_month = 217 downloads for the package `pfl`.
Capability
2 high confidencepfl-research simulates private federated learning at scale and says outright that it is not for real deployments, level with FedJAX.
- https://raw.githubusercontent.com/apple/pfl-research/HEAD/README.md recorded 2026-09-27
README: "a Python framework ... to run efficient simulations with privacy-preserving federated learning"; "multiple levels of distributed training (multiple processes, GPUs and machines)"; "local and central differential privacy"; "not intended to be used for third-party FL deployments".
Verified 2026-09-27