HisabAI - Local Business Assistant

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

HisabAI reduces repetitive paperwork for freelancers, consultants and small businesses. Users can create invoice PDFs, manage client details and extract expense information from receipts. It is designed for people who need a simple, affordable alternative to complex accounting tools and prefer to keep their financial documents on their own computers.

Demo

Code

Explore the source code, setup instructions and implementation:
https://github.com/BharatwaleJain/HisabAI

How I Built It

Built with Python, FastAPI, SQLite, Jinja2 and ReportLab, HisabAI provides invoice generation, local record storage, receipt imports and CSV exports. For AI-powered receipt extraction, it uses the open-weight Gemma 3 4B model through Ollama. A rule-based fallback is available when Ollama is disabled or unavailable. AI inference runs locally, without a remote AI provider configured.

Why Does Open Innovation Matter?

Open innovation makes AI more accessible to small businesses. Open-weight models and local inference allow users to experiment with AI without depending on paid hosted-model APIs or sending financial documents to remote AI services. Because the project is open source, developers can inspect, adapt and extend it with new models, document formats and features.

Prize Categories

Gemma — Best Use of Gemma: HisabAI uses Google's open-weight Gemma 3 4B model through Ollama for local AI-assisted receipt and invoice data extraction. This helps freelancers and small businesses extract expense details while keeping AI inference on their own machine, without relying on a hosted AI API.