Secret Encrypted Chat
Built with: Python | Django | WebRTC | MediaPipe
Secret Encrypted Chat is a multimodal interaction system that embeds a client-side encrypted chat interface behind a gesture-verification pipeline. The workflow begins in a standard restaurant ordering UI; once the user proceeds to payment, the system redirects them to a decoy “captcha” layer that initializes a real-time video stream in the browser.
Using MediaPipe’s body-landmark detection model, the video feed is processed frame-by-frame to extract skeletal keypoints. These landmarks are mapped into a normalized coordinate space and evaluated against a custom gesture classifier. Only when the user performs the required movement with sufficient confidence does the pipeline authorize the next stage.
Upon successful verification, the interface reveals a hidden communication module. All message handling — key generation, encryption, and decryption — is executed entirely on the client using browser-side cryptography. No plaintext ever touches a server, and no message history is stored or recoverable.
The result is a secure, front-end-only encrypted chat system gated behind a computer-vision gesture challenge, combining real-time pose estimation with zero-trust messaging architecture.
