OpenMined/TenSEAL

A library for doing homomorphic encryption operations on tensors

What it solves

TenSEAL allows developers to perform mathematical operations on encrypted data without needing to decrypt it first. This ensures data privacy by allowing a server or service to process information (such as performing a dot product or matrix multiplication) while the data remains encrypted throughout the computation.

How it works

It is a Python library built on top of Microsoft SEAL, implementing homomorphic encryption (HE) schemes like BFV (for integers) and CKKS (for real numbers). It provides a tensor layer that allows users to treat encrypted data as vectors or N-dimensional tensors, supporting operations like element-wise addition, subtraction, multiplication, and matrix multiplication. The core operations are implemented in C++ for efficiency, while the Python API makes the library easy to use.

Who it’s for

Developers and researchers working on privacy-preserving machine learning and secure computation, specifically those who need to perform linear algebra operations on encrypted tensors.

Highlights

  • Homomorphic Encryption Support: Supports both BFV and CKKS schemes for encrypted integers and real numbers.
  • Tensor Operations: Provides CKKSTensor and BFVTensor for N-dimensional operations, including reshape, broadcast, and transpose.
  • Linear Algebra: Built-in support for dot products and vector-matrix multiplication.
  • SEAL Integration: Provides full access to the underlying Microsoft SEAL API via tenseal.sealapi.
  • Serialization: Ability to serialize contexts, keys, and encrypted tensors for storage or transmission.

Related

  • Project
    zama-ai/tfhe-rs

    A pure Rust implementation of TFHE that enables boolean and integer arithmetics to be performed directly on encrypted data without decryption.

  • Project
    google/heir

    An MLIR-based compiler toolchain for homomorphic encryption that allows developers to write encrypted computations and deploy them across multiple HE libraries and schemes.

  • Project
    zama-ai/fhevm

    A framework for building confidential smart contracts on EVM-compatible blockchains using Fully Homomorphic Encryption (FHE) to process encrypted data on-chain.

  • Project
    QuantumKitHub/TensorKit.jl

    A Julia package for large-scale tensor computations with symmetries, optimized for building tensor network algorithms to simulate quantum many-body systems.

  • Project
    zama-ai/concrete-ml

    A privacy-preserving machine learning library that uses fully homomorphic encryption (FHE) to enable inference and training on encrypted data using scikit-learn-like APIs.