BillWise

The primary users are small business owners and accountants who currently manage invoices manually, and the key benefit is reducing compliance errors and saving time on tax preparation.

Comment

Inspiration

Small businesses struggle with invoice organization, GST categorization, expense tracking, and compliance reminders. Develop a platform that reads invoices using OCR, categorizes expenses, tracks tax deadlines, and prepares simple financial summaries. AI/ML feature: OCR-based text extraction with intelligent reading-continuity tracking.

What it does

BillWise is a web-based platform designed for small businesses and MSMEs to simplify invoice management, GST compliance, and financial tracking. The system uses OCR (Optical Character Recognition) to extract data from uploaded invoice images and PDFs, automatically categorizes expenses, tracks GST filing deadlines, and generates simple financial summaries. The primary users are small business owners and accountants who currently manage invoices manually, and the key benefit is reducing compliance errors and saving time on tax preparation.

How we built it

The proposed AI-Powered Multi-Tenant SaaS GST Invoicing and Compliance Platform (BillWise) addresses this problem through a distributed, microservices-based web platform that integrates multi-tenant merchant onboarding, role-based identity management, automated invoice generation, vendor expense tracking, dynamic GST calculation, and optical receipt digitization. The system employs a dual-database persistence tier—utilizing MySQL for ACID-compliant transactional consistency across user identities, merchant GSTINs, and role delegations, alongside MongoDB for high-throughput, schema-less storage of dynamic multi-item invoice records, tax breakdowns, and extracted optical metadata. The backend architecture leverages Spring Boot and Spring Cloud (comprising a Eureka Service Discovery Server, an intelligent API Gateway with JWT-based security filtering, an Auth & Merchant Service, and a dedicated Invoice & Compliance Service) to ensure fault isolation, high availability, and horizontal scalability. To automate document categorization and accelerate accounting workflows, the platform integrates a specialized Machine Learning Invoice Category Classifier running on a lightweight FastAPI Python microservice, which analyzes line items and transaction descriptions to automatically classify expenses into standard accounting categories. Furthermore, an integrated conversational AI component powered by Local LLM Intelligence (Ollama / AI Assistant) provides contextual assistance for tax rules, HSN code lookups, financial summaries, and invoice query resolutions. Multi-tenant governance is strictly enforced through Role-Based Access Control (RBAC)—differentiating Super Admins, Business Admins, Verified Accountants, and Viewers—fortified with Google OAuth 2.0 Single Sign-On (SSO) and automated two-factor Email OTP verification. By combining deterministic tax compliance services with AI-driven document intelligence and microservices-level security, the BillWise system significantly minimizes tax filing discrepancies, accelerates invoice lifecycle turnaround, eliminates unauthorized data manipulation, and ensures seamless, end-to-end regulatory compliance for modern enterprises.

Challenges we ran into

MySQL for ACID-compliant transactional consistency across user identities, merchant GSTINs, and role delegations, alongside MongoDB for high-throughput, schema-less storage of dynamic multi-item invoice records, tax breakdowns, and extracted optical metadata. The backend architecture leverages Spring Boot and Spring Cloud Machine Learning Invoice Category Classifier AI component powered by Local LLM Intelligence (Ollama / AI Assistant) provides contextual assistance for tax rules, HSN code lookups, financial summaries, and invoice query resolutions. Multi-tenant governance is strictly enforced through Role-Based Access Control (RBAC)—differentiating Super Admins, Business Admins, Verified Accountants, and Viewers—fortified with Google OAuth 2.0 Single Sign-On (SSO) and automated two-factor Email OTP verification

Accomplishments that we're proud of

• All backend services must be implemented in Java using Spring Boot 3.x. • All frontend interfaces must be implemented in React 18+. • Authentication must use JWT (JSON Web Token) with role-based access control. • All services must be containerised using Docker and deployable via docker-compose. • A Jenkins CI/CD pipeline must automate build, test, and deployment. • All AI/ML model inference must be exposed via a Python FastAPI REST endpoint.

What we learned

MySQL for ACID-compliant transactional consistency across user identities, merchant GSTINs, and role delegations, alongside MongoDB for high-throughput, schema-less storage of dynamic multi-item invoice records, tax breakdowns, and extracted optical metadata. The backend architecture leverages Spring Boot and Spring Cloud Machine Learning Invoice Category Classifier AI component powered by Local LLM Intelligence (Ollama / AI Assistant) provides contextual assistance for tax rules, HSN code lookups, financial summaries, and invoice query resolutions. Multi-tenant governance is strictly enforced through Role-Based Access Control (RBAC)—differentiating Super Admins, Business Admins, Verified Accountants, and Viewers—fortified with Google OAuth 2.0 Single Sign-On (SSO) and automated two-factor Email OTP verification

What's next for BillWise

updating the app based on the gstr rules and maintain the software update and a voice assisstent

Share this project:

Updates

Submission history