What do you think?

✅ 1. Microsoft Azure Face API
Real-time face comparison with similarity scores.

Threshold configurable (e.g., 80%).

Photos can be deleted after use, especially if you use their SDK or API directly.

Secure & enterprise-grade with compliance certifications.

Free tier available (limited monthly calls).

✅ Recommended for production or secure environments.

🔗 https://azure.microsoft.com/en-us/products/cognitive-services/face

✅ 2. Amazon Rekognition
Offers face comparison with similarity percentages.

You can specify a threshold (like 80%) for determining a match.

Supports temporary image usage—images can be deleted after processing if you don't upload them to S3.

AWS adheres to strict privacy and security standards.

🔗 https://aws.amazon.com/rekognition/

✅ 3. Face++ (Megvii)
Real-time face comparison with similarity score.

API deletes images after processing (stated in their documentation).

Used in many mobile and web-based apps.

Based in China, so consider data sovereignty requirements.

🔗 https://www.faceplusplus.com/face-comparing/

✅ 4. PimEyes (Search engine approach)
Not API-based, more for public facial search.

Provides similarity but not recommended for privacy-sensitive use.

May store images for indexing unless opted out.

🔗 https://pimeyes.com/

✅ 5. Luxand FaceSDK Cloud
Offers a real-time face match API.

High privacy compliance.

Face images are deleted after matching if configured.

Trusted in various biometric applications.

🔗 https://www.luxand.cloud/facematch/

🛡 Privacy Best Practice
Regardless of the provider:

Use API endpoints, not dashboard uploads.

Check privacy policies and image retention settings.

Choose providers offering on-the-fly processing (not storage).