FATE (Federated AI Technology Enabler)
FederatedAI (FATE)Industrial federated learning framework hosted by the Linux Foundation. Parties deploy their own FATE clusters, from docker-compose or KubeFATE on Kubernetes, and exchange encrypted intermediate values through a routing layer, with secure-computation protocols built on homomorphic encryption and multi-party computation, for horizontal and vertical federated learning. FATE-Flow schedules the multi-party jobs.
`pyfate` is the package FATE's own standalone deployment guide recommends installing; `fate-client` and `fate-flow` are separate components and are not declared. FATE-LLM is a separate FederatedAI repository and is treated as a SKU of this record, not a second product. The core repository was last pushed 2024-11-19. The organization is kept as `federatedai`, not LF AI & Data, because the README says "hosted by Linux Foundation" without naming LF AI & Data.
Openness
5 high confidence- license
- Apache-2.0(OSI
- source
- public(the framework and its deployment tooling)
- core features withheld
- no — foundation-hosted under a technical charter
FATE is Apache-2.0 throughout, and its repository holds one root license and no enterprise, commercial or proprietary path, so no part of it is held back for a paid tier.
- https://raw.githubusercontent.com/FederatedAI/FATE/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/FederatedAI/FATE/HEAD/README.md recorded 2026-09-27
README: "FATE is an open source project hosted by Linux Foundation. The Technical Charter sets forth the responsibilities and procedures"; FATE "implements secure computation protocols based on homomorphic encryption and multi-party computation (MPC)".
- https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/FederatedAI%2FFATE recorded 2026-09-27
Repository record for FederatedAI/FATE: license apache-2.0, not archived, not a fork, pushed 2024-11-19, 6,098 stars.
- https://ungh.cc/repos/FederatedAI/FATE/files/master recorded 2026-09-27
The 1,419-file tree of FederatedAI/FATE at `master`: 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 `pyfate`, the package FATE's standalone deployment guide recommends. Cluster deployments run from installers or container images instead, which this figure does not count.
- https://pypistats.org/api/packages/pyfate/recent recorded 2026-09-27
last_month = 74 downloads for the package `pyfate`.
- https://raw.githubusercontent.com/FederatedAI/FATE/HEAD/deploy/standalone-deploy/README.md recorded 2026-09-27
Standalone deployment guide: "Install FATE from PyPI (Recommended)", one of three single-node methods beside Docker images and host installation.
Capability
4 medium confidenceEach party runs its own FATE cluster, traffic between parties passes a routing layer that admits only listed parties, and exchanged values are protected with homomorphic encryption or MPC. That matches Flower's production reach, with no mobile or HPC path.
- https://raw.githubusercontent.com/FederatedAI/FATE/HEAD/README.md recorded 2026-09-27
README: cluster deployment by CLI, docker-compose deployment, FATE-Flow ("A multi-party secure task scheduling platform"), FATE-LLM, and "secure computation protocols based on homomorphic encryption and multi-party computation (MPC)".
- https://raw.githubusercontent.com/FederatedAI/KubeFATE/HEAD/README.md recorded 2026-09-27
KubeFATE deploys FATE with Docker Compose or Kubernetes, and manages the full life cycle of multiple FATE clusters; Kubernetes is the recommended production mode.
- https://ungh.cc/repos/FederatedAI/FATE/files/master recorded 2026-09-27
The tree ships deploy/docker-compose/docker-deploy with parties.conf, and an eggroll proxy configuration carrying route_table.json and whitelist.json.
Verified 2026-09-27