Article’s

Urban municipalities across rapidly growing cities face persistent challenges in managing civic infrastructure complaints efficiently. Conventional grievance mechanisms — telephone helplines, physical offices, and paper-based registers — are characterised by opacity, delayed response cycles, and a near-total absence of citizen feedback, leaving defects such as potholes, waterlogging, broken streetlights, and uncollected waste unresolved for extended periods. This paper presents CivicFix, a crowdsourced civic issue reporting and resolution system developed under Smart India Hackathon 2025 (Problem Statement SIH25031) for the Government of Jharkhand. The platform empowers citizens to submit geo-tagged, media-rich complaints through a Progressive Web Application (PWA) supporting multilingual voice input via the Web Speech API. A Convolutional Neural Network (CNN) engine automatically categorises incoming reports, which are routed through a four-tier administrative hierarchy — Super Admin, City Admin, Ward Admin, and Field Worker — governed by a Service Level Agreement (SLA) auto-escalation engine. Real-time status notifications are delivered through Socket.io, Twilio SMS, and email, while every administrative action is immutably recorded on a blockchain audit trail to ensure tamper-proof accountability. The system is built on a cloud-native stack comprising Next.js 14, Node.js/Express, MongoDB Atlas, Redis, and Cloudinary, deployed across Vercel and Railway. Pilot evaluation conducted across two ward offices demonstrated a 66.7 percent reduction in average issue resolution time (17.4 days to 5.8 days), a 2.3-fold increase in complaint submission volume, a 95.8 percent drop in administrative assignment latency (46 hours to 1.9 hours), and a mean citizen satisfaction score of 4.1 out of 5.0. These results confirm CivicFix as a scalable, accountable, and citizen-centric model for transforming urban governance in developing cities.