#103103rd of 168 in ML frameworks and librariesFor data and documents, data analysis
pysyft
Python library for running data-owner-approved jobs on private data
- Stars
- 10.1k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Feb 2025
- Language
- Python
9,839 on 21 Apr 202610,056 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
PySyft lets a data scientist submit Python jobs that a data owner reviews, approves and runs against private datasets, returning only approved outputs. Peers sync through cloud storage they already use, currently Google Drive, or through WebRTC channels. The client is split into packages for datasets, jobs, permissions and background services, with enclave tutorials for private LLM evaluation.
Who it is for: Researchers and orgs analysing private datasets without handing over copies
Strengths
- Private data stays on the owner's machine; scientists work on mock data until a job is approved
- Needs no new server; syncs over Google Drive, with Microsoft 365 and WebRTC also named
- Each collaborator and each job requires explicit data owner approval
- Includes tutorials for double-blind LLM evals in enclaves and LLM user log analysis
Weaknesses
- README describes only Google Drive as a working transport; others are called extensible
- Local or Jupyter use needs a Google Cloud OAuth token, about 15 minutes to set up
- Each role needs its own Google account
- Breaking API change: legacy PySyft 0.9 users must pin syft<0.10
What it needs
- Needs Python 3.10+, Google Drive, Google Cloud OAuth token (local/Jupyter)
Quick answers
Is pysyft open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Also in ML frameworks and libraries
See all 168| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |