Pulse

Real-time discovery engine surfacing emerging communities based on live momentum signals.

Pulse detects what's gaining attention right now — not what's already popular. It's built around a simple idea: the most interesting communities are often the ones just starting to spike, not the ones that are already large. Think of it like a seismograph for community activity. When a community suddenly gets more engagement than usual, Pulse notices and ranks it higher. How it works: 1. Events come in (joins, posts, views) 2. Events are weighted (a "join" has more impact than a "view") 3. Momentum is calculated (sum of weighted events in a sliding time window) 4. Rankings are cached (Redis sorted set for sub-millisecond reads) 5. Feed is served (Communities ordered by current momentum) Stack: Go + Echo for the API, PostgreSQL for reliable storage, Redis for the leaderboard, and Supabase for Auth.

Goal

Detect what's gaining attention right now using rate of change instead of absolute size.