NagrikAI – Give Every Civic Problem a Memory

• 

  landing page

• 

  landing page

• 

  Dashboard

• 

  incident and report submission

• 

  map to find report location

• 

  community driven support

Inspiration

What it does

How we built it## Inspiration

Imagine noticing a pothole on your way to college. You take a photograph and share it in a WhatsApp group or report it on social media. A few people respond, but eventually the complaint gets buried among hundreds of messages. The pothole remains, another citizen reports the same issue, and the process starts again.

We realized that the problem is not just reporting civic issues; it is the lack of a persistent record that connects the original complaint, its evidence, its history and its eventual resolution. This inspired us to build NagrikAI.

What it does

NagrikAI is an AI-powered, community-driven Civic Intelligence Platform that transforms scattered civic complaints into structured, evidence-backed incidents.

Our central concept is Civic Memory. Every reported issue receives a unique identity called a Civic Issue Digital Twin, preserving its original evidence, geographical context, status changes and update history. This helps citizens follow a problem instead of allowing it to disappear after a complaint is submitted.

Key features

• AI-powered issue analysis: Uses Google's Gemma 4 multimodal models to analyse uploaded photographs, identify civic issue categories, estimate severity and highlight potential risks.
• Location-based reporting: Citizens can report issues using GPS or an interactive map, with photo and voice-to-text support.
• Duplicate detection: PostgreSQL and PostGIS help identify potentially related incidents using geographical information.
• CBRT scoring: Our Community-Based Risk and Triage approach considers severity, public impact, community support, unresolved age and recurrence to help prioritize issues.
• Interactive civic map: Citizens can explore incidents by location, category and status using Leaflet and OpenStreetMap.
• Community participation: Users can support reports, contribute additional evidence and share updates.
• Evidence-based repair verification: Before-and-after photographs support community review of claimed repairs. Issues can be reopened when the problem remains or returns.
• Privacy and contribution tracking: Includes anonymous reporting controls, My Contributions and Saved Locations.

How we built it

We developed the frontend using React, TypeScript, Vite and Tailwind CSS. Supabase provides authentication, storage and backend services, with PostgreSQL managing incident records. We use PostGIS for geographical queries, Leaflet and OpenStreetMap for interactive mapping, and Google's Gemma 4 multimodal models for image-based civic issue analysis. Supabase Edge Functions support backend processing.

We also designed an evidence-preservation workflow so that the original report remains part of the incident's history as updates and verification evidence are added.

Challenges we ran into

One of our main challenges was converting unstructured citizen observations into structured civic incidents. We also had to consider duplicate reports, geographical matching, evidence preservation, AI-generated suggestions and community-based verification. Designing a workflow that keeps citizens involved while ensuring AI remains an assistant rather than the final decision-maker was another important challenge.

What we learned

We learned how multimodal AI, geospatial databases and community-driven workflows can work together to address real-world problems. We also understood the importance of preserving original evidence, making AI-assisted results editable and building transparent incident histories.

What's next

Our future plans include integration with municipal systems, support for more regional languages, improved risk prediction, stronger community verification and expansion to more Indian cities. NagrikAI currently does not have a direct integration with municipal ERP systems, so this remains an important future step.

Impact

NagrikAI aims to make civic reporting more organized, transparent and accountable. Instead of treating every complaint as an isolated message, we create a persistent record that communities can follow from the first observation to evidence-backed resolution.

NagrikAI – Give Every Civic Problem a Memory.

Built With

• artificial
• css
• gemma
• google
• intelligence
• leaflet.js
• learning
• machine
• openstreetmap
• postgis
• postgresql
• react
• supabase
• tailwind
• typescript
• vite