← All workCase study No. 03Mobile App + Web App + Backend03 / 17

ARSA.

AI-powered car advisory

AI-powered car advisory platform with mobile, web, expert consultations, and subscriptions.

Client: Armaan Sarkar
Industry: Automotive / AI Advisory
Engagement: Ongoing — Build + Management
Relationship: A client partnership

arsacar.in

ARSA website with car discovery and personalised advice

Visit arsacar.in

ARSA mobile app showing vehicle search and personalised car recommendations — Mobile app Vehicle discovery, designed for the palm of your hand.

IThe ground

The brief.

ARSA (Car Advisory) set out to create a complete digital platform for AI-powered car guidance in the Indian market. The founder, Armaan Sarkar of Sarcare Marketing Private Limited, wanted a technology partner who could shape the entire stack: a Flutter mobile app for car buyers, a Next.js 16 web app, a NestJS 11 backend with 39+ database entities, an AI chatbot powered by Perplexity, expert video consultations, Razorpay subscriptions, and ongoing operations. The platform brings together comprehensive vehicle search, personalised recommendations, autocomplete, model-based grouping with variant selection, expert scheduling, Free & Premium plans, ticket-based support, and an MCP connector exposing 11 read-only tools to external AI assistants—with privacy-first analytics and a strict zero-training policy.

IIThe plan

Our response.

We built a full-stack platform with three interconnected parts — a Flutter + GetX mobile app, a Next.js 16 + React 19 + Tailwind CSS v4 web app, and a NestJS 11 + Prisma + PostgreSQL backend API. As a Build + Management engagement, Edifio owns the entire tech ecosystem from architecture to deployment to day-to-day operations. The platform features a Perplexity AI-powered chatbot (sonar-pro) with 5 req/min rate limiting and smart follow-ups, an expert consultation system with scheduling, availability slots, and ratings, Razorpay subscription billing with Free & Premium tiers, usage tracking for searches and AI consultations, refund handling, ticket-based support with file attachments, Cloudinary media uploads, email via Resend, and an MCP connector exposing 11 read-only tools for Claude, ChatGPT, Cursor, and other AI assistants.

Mobile App Development
Web App Development
Backend API
AI Integration
Payment Integration
Expert Consultation System
Operations Support

IIIThe buildStrategy → craft → engineering → care

Brick by brick.

Course 01
Discovery & Architecture
Designed a 3-platform architecture for the Indian automotive market: Flutter + GetX mobile app (Android & iOS), Next.js 16 web frontend with Tailwind CSS v4 and Poppins font, and a NestJS 11 backend with Prisma ORM and PostgreSQL. Planned 39 database entities covering users, cars, experts, consultations, subscriptions, payments, support tickets, audit logs, and usage tracking.
Course 02
AI Advisory Engine
Integrated Perplexity AI (sonar-pro model) for the chatbot with 5 requests/minute rate limiting, context-aware smart follow-ups, and markdown rendering. Built comprehensive vehicle search with autocomplete, advanced filtering, car comparison engine, and personalized recommendations based on user preferences. Implemented model-based grouping with variant selection and export functionality. Added an MCP connector at /mcp exposing 11 read-only tools for external AI assistants.
Course 03
Expert Consultation & Monetization
Built expert profile management with availability slots and schedule management. Implemented video consultation booking with two consultation types: AI Chat and Human Video. Added a rating & review system for experts. Integrated Razorpay for subscription payments with Free & Premium plans, usage tracking for searches and AI consultations, automated top-up flows, and refund handling.
Course 04
Operations & Compliance
Deployed a ticket-based support system with 7 categories, admin assignment, file attachments, and reply threading. Built audit logging, dynamic app settings for support email/phone/social links, and a dashboard with key metrics. Privacy is built into the foundation through a first-party analytics posture, a strict zero-training policy for user data, 30-day server log retention, and an 18+ age policy. Deployed web to Vercel with custom .arsacar.in domain, backend to Railway, and status page at status.arsacar.in.

—A closer lookARSA in detail

ARSA website comparing vehicle specifications side by side — The details that make a confident comparison.

ARSA website with free and premium subscription plans — A simple path from discovery to expert support.

IVThe outcomeWhat took shape

What it became.

Platform Components: 3; Mobile app, web app, backend API
Database Entities: 39; Users, cars, experts, consultations, subscriptions, payments & more
MCP Tools: 11; Read-only tools exposed for external AI assistants
Engagement: Build + Management; End-to-end build, deployment & ongoing operations

Project scope and implementation details from Edifio’s case study.

Edifio understood the heart of our business and translated it into the product. The AI assistant, expert consultations, and subscription experience all came from thoughtful product conversations. They take care of the technology while we grow ARSA.

Armaan SarkarFounder, ARSA

VSpecificationBuilt on the right foundations