FacePlugin
Get in Touch
Open Source

The Open-Source Face Recognition SDK

A free, open-source face recognition toolkit for Windows and Linux — face detection, landmark detection, feature embeddings, and face similarity comparison, running entirely on-premise in Python.

Free & open sourceWindows & LinuxPython 3.9+
Woman scanning her face with a phone for 1:1 verification
Identity Verified
Face Match 99.98%
Overview

A Free Starting Point for On-Premise Face Recognition

The Open-Source Face Recognition SDK is FacePlugin's free, openly licensed toolkit for teams that want to evaluate on-premise face recognition before committing to a licensed SDK, or that need a fully open codebase they can audit and modify directly. It runs on CPU alone — a GPU is optional — and keeps every image and embedding on the machine running it.

It covers the same core building blocks as FacePlugin's licensed SDKs: face detection, landmark detection, template (embedding) extraction, and similarity comparison between two faces. It does not include the passive-liveness, anti-spoofing, or 1:N large-gallery search features that ship in the licensed Face Recognition and Face Liveness Detection SDKs — for those, see the full SDK lineup.

LanguagePython 3.9+
PlatformsWindows, Linux
HardwareCPU-only works; GPU optional
Dependency managerAnaconda recommended
Quick Start

Install and Run

  • Install Anaconda.
  • Create and activate an environment: conda create -n facesdk python=3.9 then conda activate facesdk.
  • Install dependencies: pip install -r requirements.txt.
  • Test the install: python run.py.
python
from face_recognition_sdk import FaceRecognition

face_sdk = FaceRecognition()
faces1 = face_sdk.GetImageInfo("test/1.jpg", faceMaxCount=1)
faces2 = face_sdk.GetImageInfo("test/2.png", faceMaxCount=1)

if faces1 and faces2:
    similarity = face_sdk.get_similarity(faces1[0]['embedding'], faces2[0]['embedding'])
    print(f"Similarity: {similarity}%")
    print(f"Same person: {similarity >= 75}")
FAQ

Frequently Asked Questions

Is this SDK really free to use?

Yes. It's published as open source on GitHub and doesn't require a license key or machine activation, unlike FacePlugin's commercial SDKs.

What's the license?

The repository is marked "Open Source" on GitHub. Check the repository’s own LICENSE file for the exact terms before using it in a commercial product.

How does it compare to the licensed Face Recognition SDK?

The licensed SDK adds passive liveness detection, 1:N identification at scale, native mobile builds (Android, iOS, Flutter, React Native), and vendor support — this open-source toolkit covers detection, landmarks, and 1:1 similarity comparison in Python only.

Does it send any data externally?

No. It's designed to run entirely on your own Windows or Linux machine; nothing in the documented API calls out to a remote server.

Related

Pairs Well With

Face Recognition SDK

The full licensed SDK, with liveness and 1:N search.

All SDKs

Browse every FacePlugin SDK in one place.

Face Liveness Detection SDK

Add passive anti-spoofing on top of recognition.

Need Liveness Detection, Mobile Builds, or Support?

Talk to our team about the licensed SDKs — same on-premise model, with 1:N search, native mobile platforms, and liveness detection included.