Ready-Made Real-Time Face Recognition (with Highest level of accuracy)

I need a complete, production-ready face recognition package that I can drop into an access-control workflow and run immediately in real time. The system must achieve very high accuracy and stay reliable whether it is installed indoors, outdoors, or a mix of both. It will guard multiple kinds of entry points—building entrances, gated communities, and parking lots—so robustness against variable lighting, weather, and camera angles is essential.

I am not looking to fund a research project; I want your existing, field-proven solution. An SDK or API with clear documentation is required so my team can integrate it quickly with our existing turnstile controllers and gate motors. A brief demo that proves sub-second recognition speed and low false-accept/false-reject rates on live video will be part of the handoff.

To make things smooth, please include:
• Deployment package (Docker image, install scripts, or similar)
• API or SDK documentation in English
• Recommended camera specifications and any calibration steps
• A short test report showing accuracy metrics achieved in real-time conditions

If your solution already meets these points and you can show it operating across the access points listed above, let’s talk.