Project Identity
| Team Name | Team 404 |
|---|---|
| Project Title | JanSamadhan — AI-Assisted Autonomous Civic System |
| Live URL | jansamadhan.perkkk.dev |
| Category | Smart Resource Allocation |
| SDG Targets | SDG 11 (Sustainable Cities), SDG 16 (Accountable Institutions) |
Detailed Project Quotation · Solution Challenge 2026
AI-assisted autonomous civic grievance management system for Indian urban local bodies. Built for 4,000+ ULBs, piloting in Delhi across 9 zones and 42+ complaint categories.
DOC-JS-2026-Q1
97% Complete · MVP Live
| Team Name | Team 404 |
|---|---|
| Project Title | JanSamadhan — AI-Assisted Autonomous Civic System |
| Live URL | jansamadhan.perkkk.dev |
| Category | Smart Resource Allocation |
| SDG Targets | SDG 11 (Sustainable Cities), SDG 16 (Accountable Institutions) |
Routes
39
Citizen + Authority + Admin
DB Tables
19
28 RLS policies
API Endpoints
56+
FastAPI + Next.js
Languages
10
Sarvam STT voice input
All modules are production-complete and live. The platform covers every stakeholder — citizens, field workers, authorities, and administrators.
Citizen Portal — 9 routes. Seva AI chatbot complaint filing, geo-pin with DIGIPIN, real-time tracking, heatmap, upvoting, gamification, SLA escalation, closure confirmation.
Authority Dashboard — 8 routes. Department-scoped view (RLS enforced), worker assignment, material request review, map visualization, performance reports.
Worker Mobile Web — 3 routes. Task list with SLA countdown, proof photo upload, Google Maps navigation to 4m² DIGIPIN coordinates.
Admin Control Panel — 11 routes. Platform-wide analytics, user management, CCTV surveillance dashboard, complaint moderation, wallet sync.
Public Routes — 8 routes. Landing page, public heatmap, leaderboard, legal pages. No login required.
FastAPI Backend — 35+ endpoints. Complaint intake, authority ops, worker ops, admin dashboard, CCTV proxy, WhatsApp webhook, notifications.
AI Service — 6 endpoints. YOLO v8 inference, CCTV multi-frame reliability engine, citizen corroboration, auto-ticket creation, geocoding proxy.
Database Stack — Supabase (Postgres 15 + PostGIS 3.4). 19 tables, 28 RLS policies, 14 RPC functions, 8 triggers, spatial GIST indexes, immutable audit trail.
Integrations — Gemini LLM, Mappls geocoding, Sarvam STT (10 languages), Meta WhatsApp Bot, Resend email, reCAPTCHA v2, Redis. All operational.
GCP Deployment — 3 Cloud Run services (Web, API, AI) with Cloud Build CI/CD pipeline, Artifact Registry, GCP Secret Manager.
Google Gemini
Seva chatbot · Classification
Mappls API
Geocoding · Map visualization
Sarvam STT
Voice input · 10 languages
WhatsApp (Meta)
Bot intake · Notifications
Resend
Transactional email
reCAPTCHA v2
Spam prevention
Redis (Upstash)
Cache · Sessions
Supabase
DB · Auth · Storage
Four-layer architecture designed for Indian civic infrastructure — no proprietary hardware required, and portable to government data centres.
Presentation
Desktop + mobile web interface. Hindi & English UI. No installation required. Tailwind CSS v4, Leaflet.js + Mappls maps.
Backend API
35+ REST endpoints on GCP Cloud Run. Auto-scaling, zero-downtime, 99.9% uptime SLA. JWT auth + reCAPTCHA on all public endpoints.
AI Service
Gemini classifies 42 complaint categories and 4 severity levels. YOLO v8 runs a multi-frame CCTV reliability engine with citizen corroboration.
Database
19 tables, 28 Row-Level Security policies, 14 stored RPC functions, and GIST spatial indexes.
Location
Official 10-character location code with 4m² precision. Works offline and helps prevent duplicate filings.
Deployment
3 services (Web, API, AI), CI/CD on push, secrets in GCP Secret Manager, and portable container deployment.
Security
Google OAuth + email/password with PostgreSQL RLS and controlled workflow transitions via ENUM state machine.
Voice Input
Hindi, English, Tamil, Telugu, Kannada, Malayalam, Bengali, Marathi, Gujarati, and Punjabi.
9-month development cycle from prototype to full Delhi deployment and DPCC handover.
Months 1–2.5 · Complete
Core architecture, all 4 portals, Seva chatbot, DIGIPIN integration, Supabase schema, and GCP Cloud Run deployment.
Months 2.5–5.5 · Current
1–5 ward pilot with real complaints, field worker onboarding, authority training, WhatsApp bot activation, and CCTV reliability tuning.
Months 5.5–8 · Planned
Scale to 20+ wards across 1–2 zones, onboard all 9 zones, and expand multi-department authority operations.
Month 9 · Planned
Documentation, admin training, infra transfer readiness, SLA finalization, and support transition.
Server deployment access — Permission to deploy on NIC/MeghRaj infrastructure, or approval to continue on existing GCP setup.
CCTV network integration — Permission to integrate YOLO v8 model workflows against existing Delhi CCTV streams managed by relevant departments.
Department onboarding — Designation of authority accounts for participating departments with verified government email IDs.
Primary quotation values are presented for a 1-ward pilot baseline using a 2% monthly complaint filing rate (1,860 complaints/ward/month).
SOFTWARE / CLOUD COST MODEL — PRIMARY 1-WARD PILOT BASELINE
| Service | Billing basis | Reference model | Amount (₹) |
|---|---|---|---|
| Software / Cloud CostsYear-1 Baseline | Primary 1-ward quotation baselineIncludes GCP Cloud Run, Supabase, Redis, Mappls, Gemini, Sarvam STT, WhatsApp Cloud API, and email operations. | ₹19,750 / month × 12 months | ₹2,37,000 |
| Software/cloud reference — 1 ward / month | ₹19,750 | ||
| Software/cloud reference — 1 ward / year | ₹2,37,000 | ||
FULL YEAR-1 QUOTATION — PRIMARY 1-WARD PILOT (EXCL. GST)
| Line item | Calculation basis | Amount (₹) |
|---|---|---|
| Prototype & Testing CostMVP development + QA cycle | Full-stack implementation across web, API, and AI services, including end-to-end user-role integration testing. | ₹4,50,000 |
| Machinery / HardwareNo proprietary hardware required | Cloud infrastructure provided by government data centres or GCP | ₹0 |
| Team Stipends5 member implementation team | 5 members × 9 months × ₹22,000 / month | ₹9,90,000 |
| Execution & DeploymentDeployment readiness and field execution | Fixed execution allocation covering survey, commissioning, deployment configuration, and handover readiness activities. | ₹50,000 |
| Cloud & Connectivity (Year 1)Primary 1-ward pilot software/cloud model | ₹19,750 / month × 12 months × 1 ward | ₹2,37,000 |
| Dashboard & App DevAPI toolkit + third-party development costs | Gemini, Sarvam STT, WhatsApp API, Mappls, and developer toolkit allocation | ₹60,000 |
| Contingency Fund (10%)Risk reserve for internal implementation and scale-up contingencies | Approved contingency reserve aligned to internal project-cost coverage in the quotation sheet. | ₹1,42,700 |
| Total Estimated Cost — Year 1 · 1-Ward Pilot (Excl. GST) | ₹19,29,700 | |
| Per-grievance infrastructure cost (1-ward, software/cloud only) | ₹10.62 | |
| Per-grievance full-project cost (1-ward, Excl. GST) | ₹86.46 | |
| Complaint volume basis (1 ward × 12 months) | 22,320 | |
| GST is excluded from totals above and applied separately at invoicing as per applicable tax rules. | ||
Drag the slider to project Year-1 software/cloud cost from the same 1-ward baseline used in this quotation.
Monthly infra cost
₹19,750
1 ward
Annual infra cost
₹2.37 L
cloud services only
Complaints / year
22K
at 2% filing rate
Per-grievance cost
₹10.62
infrastructure only
| Item | Scope | What's included | Annual Cost (₹) |
|---|---|---|---|
| Cloud Infrastructure | 1-ward scale (expandable up to 7 wards) | GCP Cloud Run, Supabase Pro, Redis, domain renewal, and SSL baseline operations. | ₹1,60,000 |
| Developer Support | Part-time retainer (3 days/week) | Security patches, dependency updates, bug fixes, and admin support. | ₹1,20,000 |
| API Renewals | Third-party service continuity | Gemini, Sarvam STT, WhatsApp Cloud API, Mappls, and transactional email stack. | ₹36,000 |
| Annual Maintenance Total (Post Year-1) | ₹3,16,000 | ||
| Annual maintenance cloud baseline corresponds to approximately ₹13,333 per month. | |||
Cloud infrastructure operations — Monitoring, uptime management, and auto-scaling configuration.
Security and dependencies — Security patches, dependency upgrades, and Supabase schema migrations.
AI lifecycle management — Gemini version upgrades and YOLO retraining for Delhi-specific CCTV patterns.
WhatsApp policy compliance — Meta policy alignment and integration maintenance.
Dedicated support retainer — One dedicated developer for fixes, requests, and SLA reporting.
JanSamadhan is purpose-built for urban local bodies that central systems and generic CRM vendors underserve, with a strong per-grievance cost advantage.
CPGRAMS cost / grievance
₹83–250
National scale estimate · DARPG budget
JanSamadhan infra-only / grievance
₹10.62
1-ward Year-1 software/cloud baseline
Cost advantage
8–24×
Based on infra-only per-grievance baseline
Traditional-vendor delta
60–70%
Lower estimated cost vs conventional builds
| Capability | CPGRAMS | Salesforce Gov Cloud | JanSamadhan |
|---|---|---|---|
| AI complaint classification | No | Partial | Yes — 42 categories, 4 severity levels |
| Spatial deduplication | No | No | Yes — PostGIS 20m radius |
| Field worker dispatch + DIGIPIN navigation | No | No | Yes — 4m² precision |
| WhatsApp bot intake | No | No | Yes — full flow |
| CCTV auto-detection | No | No | Yes — YOLO v8 + reliability engine |
| Community severity escalation | No | No | Yes — upvote to severity bump |
| Indian language voice input | No | No | Yes — 10 languages (Sarvam STT) |
| Per-grievance cost basis | ₹83–250 | Typically license-heavy | ₹10.62 infra-only · ₹86.46 full-project |