MudraGen — PMMY Loan Project Generator

A Next.js + Firebase SaaS that generates PMMY (Pradhan Mantri MUDRA Yojana) loan project reports for Indian entrepreneurs. Built by Anoint Tech on a web stack designed for Indian compliance work.

The client

MudraGen serves Indian entrepreneurs applying for loans under the Pradhan Mantri MUDRA Yojana — the central government's microfinance scheme administered through banks and NBFCs for Shishu, Kishore, and Tarun loans up to ₹10 lakh. Every applicant needs a project report: a structured document covering the business description, financial projections, capital expenditure, working-capital requirements, and repayment plan. Preparing this by hand is a multi-day exercise that most first-time applicants outsource to consultants, often at a cost that defeats the purpose of applying for a small-ticket loan.

MudraGen turns the project-report problem into a guided web workflow. An entrepreneur answers a structured questionnaire about their proposed business; the platform assembles a bank-submittable project report with the right structure, the right financial tables, and the right supporting narrative.

The challenge

The core product problem was encoding the domain. PMMY project reports are not free-form documents — they have conventional sections that loan officers expect to see in a particular order, with specific tables and computations. The tolerable variation is narrow, and getting it wrong means the applicant's report gets questioned at the branch. We had to capture that convention in a rule-driven document generator that works across business categories (retail, manufacturing, services, agri-allied) each with their own sub-conventions.

The second challenge was the cost model. PMMY applicants are price-sensitive by definition — someone applying for a ₹50,000 Shishu loan is not going to pay ₹10,000 for a report. The platform needed to run cheaply enough per-generated-report that a low subscription price was sustainable. This directly shaped the backend architecture.

The third challenge was Indian payments and billing. Razorpay subscriptions with UPI, card, and netbanking were table stakes; so was a clean webhook pipeline for subscription lifecycle and refund handling.

What we built

MudraGen is a Next.js 14 web application on the App Router, backed by Firestore, Firebase Auth, and Razorpay. The questionnaire is a multi-step form with inline validation, autosave on every field change, and a resume-where-you-left-off flow. The report generator runs server-side in Cloud Run, consumes the questionnaire answers, renders a project-report document in HTML, and ships a PDF back to the user. The PDF renderer uses a deterministic layout engine so the same inputs always produce the same output — critical when a bank loan officer wants to verify the report against the applicant's claims.

Payments use Razorpay's standard checkout integrated via their official Next.js patterns. Subscription state is synced from Razorpay webhooks into Firestore, with idempotent handlers that survive duplicate or out-of-order delivery. Users can upgrade, downgrade, or cancel from the app; their report history persists through all of it.

SEO was part of the brief from day one. The public marketing surfaces (landing page, FAQs, PMMY educational content) are server-rendered with proper metadata, structured data, and a sitemap that grows as new content pages ship.

Technology stack

  • Front end: Next.js 14 (App Router), React 18, TypeScript, Tailwind CSS.
  • Back end: Node.js on Cloud Run, Next.js API routes, Firebase Cloud Functions for lightweight triggers.
  • Data: Cloud Firestore (primary), Cloud Storage (generated PDFs), BigQuery (analytics).
  • Auth: Firebase Auth with email/password and Google sign-in.
  • Payments: Razorpay subscriptions (UPI, card, netbanking), webhook verification, refund and cancellation handling.
  • PDF generation: Headless Chrome on Cloud Run with a deterministic layout template.
  • Infrastructure: Cloud Run, Firebase Hosting, Cloud Build, Secret Manager.
  • Observability: Cloud Logging, Cloud Trace, Firebase Crashlytics for the web client.

Outcome

MudraGen runs in production at mudragen.com. Operating cost per generated report stayed within the envelope that makes low-priced subscription tiers viable — the architectural choice to lean heavily on Firestore, Cloud Run scale-to-zero, and Firebase's free-tier services paid off. For Anoint Tech, MudraGen is the canonical example of how we ship an Indian-market SaaS on the Next.js + Firebase stack: compliance-aware, Razorpay-integrated, and cost-tuned for price-sensitive users.

Related

Building an Indian SaaS?

Next.js + Firebase + Razorpay is our default stack. We scope MVPs at fixed price with week-by-week milestones. Tell us what you want to ship.

Start a conversation →