# google/fully-homomorphic-encryption

Homomorphic Encryption demos

Repository: https://github.com/google/fully-homomorphic-encryption
Canonical: https://ross.abutalabs.com/products/fully-homomorphic-encryption
Language: Starlark
License: Apache-2.0
License Family: permissive
Topics: fully-homomorphic-encryption, privacy, cryptography, homomorphic-encryption, fhe, compiler, encryption
Last push: 2026-08-16T17:37:09+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 62, longevity 100
- inputs: {"age_days": 1946, "days_push": 17, "days_rel": 40, "gap_med": 437, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3708, forks 274 (observed 2026-08-28T04:08:15.170860+00:00)

## What it is
Google's repository of demos for fully homomorphic encryption (FHE), originally a C++ transpiler and now showcasing the HEIR MLIR-based FHE compiler toolchain and Jaxite backend. FHE allows computation on encrypted data, enabling privacy-preserving machine learning inference without decrypting inputs.

## Use cases
- run machine learning inference on encrypted data
- convert existing models to FHE-compatible versions
- learn how fully homomorphic encryption works with demos
- process sensitive user data without server-side exposure
- experiment with privacy-preserving computation toolchains

## When to choose
- you want working demos of privacy-preserving encrypted inference
- you need to explore Google's FHE ecosystem including HEIR and Jaxite
- you want to compile models to run on encrypted data without cryptography expertise

## When to avoid
- you need a production-ready FHE library rather than demos
- you need high-performance general-purpose encryption rather than encrypted computation
- you want the original transpiler, which is archived

## Facets
- artifact type: library
- maturity: maintenance
- function: compiler, cryptography, security, privacy, machine-learning
- domain: privacy, machine-learning, compilers, security
- platform: cpp, python, cross-platform
- tags: fully-homomorphic-encryption, fhe, mlir, transpiler, private-inference, demos, cryptography

## Member repositories
- google/fully-homomorphic-encryption (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:15.170860+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:30:57.010423+00:00, confidence not recorded.
  - readme: https://github.com/google/fully-homomorphic-encryption (fetched 2026-08-28T04:08:15.170860+00:00, sha 3796cf485021)
- Data as of 2026-08-30T08:39:29.467469+00:00.