# deepinsight/insightface

State-of-the-art 2D and 3D Face Analysis Project

Repository: https://github.com/deepinsight/insightface
Canonical: https://ross.abutalabs.com/products/insightface
Homepage: https://insightface.ai
Language: Python
License Family: other
Topics: face-recognition, face-detection, mxnet, face-alignment, age-estimation, arcface, retinaface, pytorch, oneflow, paddlepaddle
Last push: 2026-07-27T03:22:16+00:00

## Health v2 (maintenance only)
Score: 65/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 94, release rhythm 8, longevity 100
- inputs: {"age_days": 3289, "days_push": 37, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 29580, forks 6076 (observed 2026-08-28T04:11:53.561535+00:00)

## What it is
InsightFace is an open-source 2D and 3D face analysis project providing state-of-the-art face detection, recognition, alignment, and face swap models (ArcFace, RetinaFace, InSwapper) with a Python library and a self-hosted face recognition server. It includes a cross-platform GUI evaluation studio and REST API server that runs local ONNX Runtime inference.

## Use cases
- detect and recognize faces in images
- extract face embeddings for identity verification
- build a self-hosted face recognition server like AWS Rekognition
- swap faces in photos or live streams
- estimate age and align faces
- benchmark face recognition models
- search large image databases for a person

## When to choose
- you need state-of-the-art, NIST FRVT-ranked face detection or recognition models
- you want a self-hosted alternative to cloud face recognition APIs
- you need research-grade face analysis models with Python/ONNX inference
- you need face swap or liveness detection capabilities

## When to avoid
- you need commercial deployment without licensing - pretrained models are non-commercial research only
- you need general object detection or non-face computer vision
- you want a fully open commercial license for the inswapper models
- you need a managed cloud service rather than self-hosting

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, computer-vision, image-processing, sdk, http-server
- domain: computer-vision, artificial-intelligence, deep-learning, image-processing
- platform: python, cross-platform, windows
- tags: face-recognition, face-detection, face-swap, arcface, retinaface, onnx, face-alignment, age-estimation, deepfake-detection, self-hosted, linux, macos, docker, gpu

## Member repositories
- deepinsight/insightface (main) score 65

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:53.561535+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 2026-08-29T16:52:54.998094+00:00, confidence not recorded.
  - readme: https://github.com/deepinsight/insightface (fetched 2026-08-28T04:11:53.561535+00:00, sha 66db2db0700d)
  - homepage: https://insightface.ai (fetched 2026-08-29T07:50:06.678318+00:00, sha bfca1a69e2be)
  - site_page: https://www.insightface.ai/about (fetched 2026-08-29T07:50:06.690744+00:00, sha e53bf4b49e78)
  - registry_pypi: https://pypi.org/pypi/insightface/json (fetched 2026-08-29T07:50:06.693046+00:00, sha 82cb500d32e0)
- Data as of 2026-08-30T08:39:29.467469+00:00.