MERN + AI
Full-Stack MERN + AI
CivilEye
Full-stack MERN complaint management platform with role-based workflows for User, Investigator, Admin, and Super Admin. Added AI-assisted complaint triage with Groq to suggest complaint category, priority, and reasoning during intake while keeping the core case workflow secure and non-blocking.
Problem
Manual complaint handling was slow, opaque, and
difficult to audit. The AI enhancement improves intake by helping users pick
better categories and giving admins/investigators an early priority signal.
System Design
MERN architecture with separate frontend/backend apps,
role-protected complaint flows, and a protected triage endpoint that calls
Groq, validates JSON, and stores AI metadata on complaint records.
Security
JWT auth, role-based protected routes, bcryptjs
hashing, OTP verification, environment-based secrets, Multer upload handling,
and a backend-only Groq API key that is never exposed to the frontend.
AI / Intelligence
Groq API with llama-3.3-70b-versatile returns structured
JSON for suggested category, priority, and reasoning; if the AI call fails,
the complaint workflow continues normally.
Highlights
- Built a full-stack MERN crime reporting and case management system.
- Implemented role-based workflows for users, admins, investigators, and super admin.
- Added AI-assisted complaint triage with Groq for category and priority suggestions.
- Improved case workflow with public investigation progress and internal notes/evidence separation.
- Added secure authentication with JWT, OTP verification, protected routes, and environment-based secrets.