Open-Source-Face-Recognition-SDK

Why we included this project

Teams working on access control, attendance, or identity verification get a useful starting point here: the SDK wraps the full face pipeline (detection, landmark extraction, embedding generation, and similarity comparison) in a small Python API that runs entirely on-premise. Everything executes locally on Windows or Linux, so no face data leaves the machine, which matters for privacy-sensitive deployments. It runs on CPU-only systems with GPU as an option, and the quick-start examples show how to detect multiple faces in an image and compare two embeddings against a configurable similarity threshold. That makes it a reasonable base for prototyping biometric login or surveillance workflows before moving to a commercial offering with liveness detection and broader platform support.