JanMitra AI: Reshaping Municipal Citizen Complaint Handling Systems with Multimodal AI

JanMitra AI is a municipal complaint handling and predictive analysis platform for Uttar Pradesh, India, built on Google Gemini. It integrates multimodal AI (visual, voice), GIS geolocation, intelligent routing, and other features to transform traditional municipal management into a transparent and efficient modern governance solution. The project is open-source, maintained by theabhishek4u, and was released on GitHub on June 1, 2026. Its core goals are to lower the threshold for citizen participation and enhance administrative accountability.

JanMitra AI (meaning "Friend of the People") is designed for the municipal departments of Uttar Pradesh, India, to address the "black box" issues in traditional public administration (opaque processes, low efficiency in complaint handling). Its core philosophy is to make government services accessible, reduce technical barriers through multimodal inputs (voice, images), and achieve transparent governance and build citizen trust.

Core features include: 1. Multimodal AI visual scanning (Gemini Vision identifies image-based issues); 2. Speech-to-text (supports Hindi, Urdu, Hinglish, and other dialects); 3. Real-time interactive dashboard (visualizes complaint handling processes); 4. GIS geolocation (precisely locates issues); 5. Holographic tracking timeline (transparent status of the entire process). The tech stack includes Next.js, React, Tailwind CSS, Google Gemini, Leaflet.js, etc.

The system improves efficiency through automation: automatically routes complaints to relevant departments (municipal, water supply, electricity, etc.); merges duplicate work orders to avoid resource waste; an autonomous node agent follow-up engine performs dynamic audits and SLA tracking within a 5-day cycle (e.g., classify complaints on day 1, push reminders to officials on day 2, verify solutions on day 5) to ensure complaints are not delayed.

The system uses spatial clustering algorithms to identify regional issues: when there are ≥2 unresolved complaints of the same category in the same area, it is marked as a "hotspot", priority is raised, and it bubbles to the top of the officials' queue; after some complaints are resolved, the cluster shrinks and priority is restored. This mechanism helps management departments identify systemic failures and optimize resource allocation.

JanMitra AI provides an important direction for civic technology: lowering participation thresholds, enhancing transparency, intelligent resource allocation, and automated accountability. For developers, it demonstrates the potential of combining multimodal AI with GIS to solve practical problems. This project is a model of AI applied to social welfare and is worth attention and reference.