Bridging the gap between citizens and authorities with AI, transparency, and real-time accountability.
Build an AI-powered civic infrastructure safety & accountability platform that enables citizens to report public infrastructure issues, allows authorities to triage and assign tasks to engineers, and creates a transparent, data-driven ecosystem for resolving civic problems like road damage, bridge cracks, water supply issues, and more.
India's public infrastructure — roads, bridges, water supply, and public buildings — faces a massive accountability gap. Citizens struggle to report issues, authorities lack real-time visibility, and resolution is slow, opaque, and disconnected. JanaVaani solves this by putting AI, geolocation, and transparency at the center of civic problem-solving.
JanaVaani is a three-portal intelligent ecosystem that connects Citizens → Authorities → Engineers in a seamless, AI-augmented workflow:
| Portal | Who | What They Do |
|---|---|---|
| 🏠 Citizen | Public | Report issues via photo/voice, track on live map, upvote problems |
| 🏛️ Authority | Govt Officials | Triage reports with AI risk scores, assign engineers, monitor analytics |
| 🔧 Engineer | Field Staff | Receive tasks, upload inspection evidence, mark resolution |
flowchart LR
A[🧑 Citizen<br/>Reports Issue] -->|📸 Photo + 📍 GPS| B[🤖 AI Engine]
B --> C{🔍 Authentic?}
C -->|❌ No| D[🚫 Reject Fake]
C -->|✅ Yes| E[🗺️ Map + Leaderboard]
E --> F[🏛️ Authority<br/>Dashboard]
F --> G{⚠️ Escalate?}
G -->|Yes| H[🚨 Auto-Escalation]
G -->|No| I[🔧 Assign Engineer]
I --> J[✅ Task Completed]
J --> K[📊 Accountability Score]
- Image Classification — Hugging Face models auto-detect issue type (road crack, water leak, etc.)
- Duplicate Detection — CLIP ViT embeddings identify similar/duplicate reports automatically
- Authenticity Verification — AI + EXIF metadata block fake or tampered images
- Voice-to-Text Reporting — Speak your complaint; AI parses and structures it
- Citizen Portal — Anonymous reporting, live map view, risk score tracking, gamified leaderboard
- Authority Portal — Real-time admin dashboard for PWD/municipal/state officials
- Engineer Portal — Field task management with evidence upload and status updates
- Auto Geolocation — GPS + reverse geocoding via OpenStreetMap
- Interactive Leaflet Maps — Clustered, color-coded issue visualization
- Risk Scoring — Dynamic severity assessment with priority mapping
- Karnataka District Mapping — Tailored for state-level infrastructure tracking
- Real-time Dashboard — Live metrics, trend analysis, and resolution timelines
- Gamified Leaderboard — Citizen engagement via upvotes and reputation scores
- Accountability Scoring — Track authority and engineer performance over time
- Auto-Escalation — Unresolved issues automatically escalate based on priority & time
graph TB
subgraph "Frontend"
A[Next.js 16 App Router]
B[React 19 + Tailwind CSS]
C[Leaflet Maps]
D[Recharts Analytics]
end
subgraph "Backend & AI"
E[Next.js API Routes]
F[Supabase PostgreSQL]
G[Hugging Face Inference]
H[CLIP ViT Embeddings]
I[Gemini AI Parsing]
end
subgraph "Data Layer"
F
J[Row Level Security]
K[Vector Embeddings Table]
end
A --> E
E --> F
E --> G
E --> H
E --> I
B --> C
B --> D
F --> J
F --> K
| Layer | Technology | Purpose |
|---|---|---|
| Framework | Next.js 16.2.6 + React 19.2.5 | SSR, App Router, modern React |
| Styling | Tailwind CSS + Glass Morphism | Responsive, modern UI |
| Database | Supabase (PostgreSQL) | Real-time DB, auth, storage |
| Security | Row Level Security (RLS) + Zod | Granular access + input validation |
| AI/ML | Hugging Face Inference API, CLIP ViT, Gemini | Classification, embeddings, parsing |
| Maps | Leaflet + React-Leaflet | Interactive geolocation visualization |
| Charts | Recharts | Analytics dashboards |
| Forms | React Hook Form + Zod | Type-safe validation |
Try the full end-to-end workflow:
admin@pwd.karnataka.gov.in |
|
| Password | password123 |
| Capabilities | Review reports, update statuses, assign tasks, view analytics |
engineer@pwd.karnataka.gov.in |
|
| Password | password123 |
| Capabilities | View assigned tasks, upload evidence, mark COMPLETED |
No login required! Report issues anonymously with photo + location.
# 1. Clone the repo
git clone https://github.com/Code-odyssey-hackathon/team-25.git
cd team-25
# 2. Install dependencies
npm install
# 3. Configure environment
cp .env.example .env
# Edit .env with your Supabase URL, Anon Key, and Hugging Face API key
# 4. Run locally
npm run dev
# Open http://localhost:3000erDiagram
USERS ||--o{ REPORTS : submits
REPORTS ||--o{ REPORT_EMBEDDINGS : has
REPORTS ||--o{ MASTER_TICKETS : generates
AUTHORITIES ||--o{ MASTER_TICKETS : manages
ENGINEERS ||--o{ MASTER_TICKETS : assigned_to
REPORTS ||--o{ UPVOTES : receives
USERS {
uuid id PK
string phone
int reputation_score
}
REPORTS {
uuid id PK
string issue_type
string status
float lat
float lng
string image_url
int risk_score
}
REPORT_EMBEDDINGS {
uuid report_id PK
vector embedding
}
MASTER_TICKETS {
uuid id PK
uuid report_id FK
uuid engineer_id FK
string priority
string status
timestamp deadline
}
- ✅ Row Level Security (RLS) — Users only see authorized data
- ✅ AI Fraud Prevention — Fake/tampered images blocked at pixel level
- ✅ JWT Sessions — Secure auth via Supabase
- ✅ SQL Injection Safe — Parameterized queries + Zod validation
| Metric | Target |
|---|---|
| 🕒 Report-to-Resolution Time | Reduce by 60% via auto-triage & assignment |
| 🎯 Fake Report Detection | 90%+ accuracy with AI + EXIF validation |
| 📍 Geographic Coverage | District-level granularity for Karnataka |
| 🏆 Citizen Engagement | Gamified leaderboard drives repeat participation |
JanaVaani transforms civic complaint boxes into real-time, AI-driven infrastructure command centers.
Code Odyssey Hackathon 2026
| Role | Focus |
| Full-Stack Engineering | Next.js, Supabase, API Architecture |
| AI & ML Integration | Hugging Face, CLIP, Gemini, Embeddings |
| UI/UX & Maps | Tailwind, Leaflet, Dashboard Design |
| Database & Security | PostgreSQL, RLS, Schema Design |
