I'm the creator of this project. I built ECHOLOCK for the recent Cyber AI Hackathon 2025.
The Problem: I was frustrated by the lag in threat intelligence. When one system detects a new phishing site, it can take hours or days for that information to protect everyone else.
The Solution: ECHOLOCK is a hybrid detection system, but its main feature is a federated network. When any node in the network detects a new, high-confidence threat, it uses Redis Pub/Sub to instantly publish that threat's fingerprint to all other nodes. This means the entire network gets immunity in seconds from a threat only one node has seen.
The backend is built with Flask, and the AI model is a LinearSVC which I chose for its speed and small footprint (91% accuracy at 45MB). The README has more technical details and architecture diagrams.
The project is open-source, and I'd love to get any technical feedback you have on the architecture or the federated approach.
Hi HN,
I'm the creator of this project. I built ECHOLOCK for the recent Cyber AI Hackathon 2025.
The Problem: I was frustrated by the lag in threat intelligence. When one system detects a new phishing site, it can take hours or days for that information to protect everyone else.
The Solution: ECHOLOCK is a hybrid detection system, but its main feature is a federated network. When any node in the network detects a new, high-confidence threat, it uses Redis Pub/Sub to instantly publish that threat's fingerprint to all other nodes. This means the entire network gets immunity in seconds from a threat only one node has seen.
The backend is built with Flask, and the AI model is a LinearSVC which I chose for its speed and small footprint (91% accuracy at 45MB). The README has more technical details and architecture diagrams.
The project is open-source, and I'd love to get any technical feedback you have on the architecture or the federated approach.
Thanks!