Hey Bubblers! ![]()
We all know the struggle: you want to add Face Recognition or Biometric Login to your app, but you get stuck with complex setups like Amazon Rekognition, Azure Face API, or Google Cloud.
And then⦠the bills start coming in for every single API call. ![]()
I built FaceOS to solve this. Itās a powerful AI Vision engine that runs 100% locally in your userās browser using TensorFlow.js.
This means:
-
Zero API Costs: Scan 1 face or 1 million faces. It costs you $0.
-
Privacy First: Images are processed in the browser memory. They never leave the device unless you explicitly save a snapshot. GDPR compliance made easy.
-
Real-Time: No network latency. It works instantly.
Iāve released two editions to suit different needs:
FaceOS - Pro Edition (The Biometric Suite)
This is a full-blown security and analytics tool. It converts faces into mathematical ādescriptorsā that you can save to your Bubble database to identify users.
Pro Exclusive Features:
-
Face Recognition (Login): Compare the live camera feed against your database of users. If it finds a match, you get the Userās Name and Confidence Score. Perfect for passwordless login or clock-in systems. -
Emotion AI: Detects 7 emotions in real-time (Happy, Sad, Angry, Surprised, etc.). Great for customer satisfaction apps. -
Demographics: Instantly estimates Age and Gender (e.g., āMale, approx 30 years oldā). Useful for smart retail kiosks. -
Security Snapshots: Capture a Base64 image of the user the exact moment they are recognized for your audit logs. -
Mobile Ready: Includes actions to instantly switch between Front (Selfie) and Back (Environment) cameras.
Get the Pro Version here: PRO VERSION
FaceOS - Free Edition (The Detector)
I wanted to give the community a robust tool for basic detection without any barriers.
What can the Free version do?
-
Face Detection: Draws a bounding box around every face in the frame.
-
People Counter: Returns the number of people currently looking at the camera (State: Face Count).
-
Auto-Start: Wake up the camera immediately or control it via workflows.
-
Smart Triggers: Events for Face Detected and Face Lost.
Get the Free Version here: FREE VERSION
How does the āLoginā work?
Itās simple math, not magic.
-
Enrollment: When a user signs up, the plugin scans their face and outputs a JSON text string (Descriptor). You save this to the Userās thing.
-
Login: When a user returns, you load that list of Descriptors into the plugin. The plugin compares the live video to the list and tells you: āItās User #123 with 98% confidence.ā
No data is sent to me, Amazon, or Google. You own the data.
Let me know if you have any questions or feature requests! ![]()
Happy Bubbling!