CurationGate — Invite-Based Community Content Moderation

Stop AI spam before it joins — invite trees that self-regulate quality

CurationGate uses hierarchical invite codes with decay mechanics to create natural friction against bot networks flooding Discord and Slack communities. When invitees post spam or AI slop, their inviters lose invite power across generations, making quality gatekeeping self-enforcing. Communities see 70-90% reduction in moderation overhead within the first month by preventing bad actors at the door instead of cleaning up after them.

Key Benefits:

- Hierarchical invite trees with HMAC-SHA256 verification create cryptographic accountability chains — each invite tier automatically loses power when downstream users trigger spam detection

- PostgreSQL-backed reputation scoring with Redis rate limiting catches coordinated bot networks by tracking invite pattern anomalies and IP clustering across the tree structure

- Real-time analytics dashboard shows exactly which invite branches are producing quality members versus spam sources, letting moderators prune bad actors at the root with one click

MVP Scope: Build core invite system with hierarchical tree structure, power score decay mechanism that reduces invite value with each generation tier, basic spam detection using rate limiting and IP/domain tracking, and simple user reputation scoring. Include invite code generation with metadata, validation/redemption logic, and basic analytics dashboard to monitor invite distribution and redemption patterns.

Tech Stack: PostgreSQL, Redis, UUID v7, HMAC-SHA256, Node.js/Python backend, React frontend, Webhook callbacks

Components:

- Hierarchical Invite Tree Engine

- Power Score Decay Calculator

- Spam Detection & Rate Limiter

- User Reputation System

- Analytics & Monitoring Dashboard

Quality assessment: Strong concept with novel decay mechanics addressing real spam problems in communities, solid tech stack and clear value proposition, but lacks depth on implementation details (how decay formula works, edge cases, actual validation of 70-90% claim) and market validation beyond assertion.