DEV Community: Sai Nandi

DEVLEDGER

Sai Nandi — Tue, 13 Oct 2026 15:37:50 +0000

DEVLEDGER — Your Private AI Finance Assistant 💸🤖

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

What I Built

DEVLEDGER is an AI-powered finance management application designed to help freelancers, independent contractors, and small business owners manage their invoices, receipts, and expenses more efficiently.

Freelancers often spend valuable time organizing receipts, tracking expenses, and preparing financial reports. At the same time, uploading sensitive financial documents to third-party services can raise privacy concerns.

DEVLEDGER aims to make financial management simpler by bringing document organization, expense tracking, and AI-powered insights together in one place.

✨ Key Features

📄 Invoice & Receipt Management: Organize financial documents in one place.
🤖 AI-Powered Extraction: Extract important information from invoices and receipts.
📊 Expense Analytics: Understand spending patterns through dashboards and visualizations.
🔍 Smart Financial Queries: Ask questions about expenses and financial records.
📁 Transaction Management: Track and categorize income and expenses.
🔒 Privacy-First Approach: Designed around local AI processing to help keep sensitive financial information on the user's device.
💼 Side-Hustle Potential: A practical productivity tool for freelancers and small businesses that want to spend less time on manual bookkeeping.

The goal is to turn repetitive financial tasks into a simpler, smarter workflow.

Demo

A live demo or video demonstration will be added here when available.

Code

Explore the project and its source code:

🔗 GitHub: https://github.com/Harshini-Nandi/DEVLEDGER

The project is designed to combine a modern web application with AI-powered document processing and financial management capabilities.

Contributions, suggestions, and feedback are welcome!

How I Built It

I approached DEVLEDGER as a practical application of open-source AI to a real-world productivity problem.

The project brings together a modern web interface, backend services, financial data management, and AI-assisted document processing.

The intended technology stack includes:

React + Vite: For building a responsive and interactive frontend.
Tailwind CSS: For a clean, modern interface.
Python + FastAPI: For backend APIs and document-processing workflows.
SQLite: For lightweight, locally stored financial records.
Ollama + an open-weight language model: For local AI inference.
PyMuPDF + Tesseract OCR: For extracting text from PDFs and scanned receipts.
Recharts: For financial analytics and visualizations.

The AI component is intended to help extract structured information from financial documents and make stored financial records easier to understand.

Rather than relying entirely on a hosted AI API, the project explores running an open-weight model locally. This approach provides greater control over model selection, data processing, and the application's future development.

The technology list above describes the intended stack; components should be listed as implemented only after they have been integrated and tested.

Why Does Open Innovation Matter?

Open innovation is central to the idea behind DEVLEDGER.

Financial documents can contain sensitive information, including vendor details, transaction amounts, and business expenses. Sending this information to an external AI service is not always desirable.

Using a locally hosted open-weight model offers several potential advantages:

🔐 Privacy and Control: Local inference can keep document contents on the user's machine instead of sending them to a third-party AI API.

💰 Lower Recurring AI Costs: Once a suitable model is installed, users can avoid per-request charges from hosted AI providers, although hardware, electricity, and maintenance still have costs.

🔄 Model Flexibility: Developers can experiment with compatible models and choose the best balance of speed, accuracy, and resource requirements.

🛠️ Customization: Open-weight models and open-source tools make it possible to adapt the workflow for specific document formats and business needs.

🌱 Reduced Vendor Lock-In: The application can be designed around replaceable AI components instead of depending entirely on one provider's API or pricing.

For a small freelancer or business, these advantages could make an AI-assisted finance tool more accessible and give users greater control over their information.

Open innovation is not simply about using a free model. It is about giving developers and users more flexibility over how their tools work, where their data goes, and how their applications evolve.

My Agent Session

I used AI-assisted development to help build and refine DEVLEDGER.

An agent session or development walkthrough can be added here to show the implementation process, architectural decisions, and debugging workflow.

Prize Categories

Primary category: Open-Source AI / Side Hustle

DEVLEDGER focuses on applying open-weight AI to a practical business problem, with an emphasis on privacy, affordability, and useful financial automation.

Building DEVLEDGER has been an opportunity to explore how open-source AI can move beyond experimentation and become part of a practical tool that people could use in their everyday work.

I'd love to hear your feedback and suggestions! 🚀

wildbuddy

Sai Nandi — Sat, 10 Oct 2026 17:45:09 +0000

*# 🌿 WildBuddy — Your Offline AI Nature Companion

This is my submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass.

🌱 What I Built

WildBuddy is an offline-first, open-source AI nature companion designed to help people spend less time staring at screens and more time exploring the outdoors.

In a world where many apps encourage endless scrolling, WildBuddy takes a different approach: it uses technology to encourage people to disconnect from their screens and reconnect with nature.

Here's what WildBuddy offers:

🌿 AI Nature Identification: Identify leaves, flowers, plants, birds, and insects using locally running AI vision models.
🚶 Outdoor Adventures: Complete nature missions that encourage users to explore their surroundings and observe biodiversity.
📸 Nature Journal: Save observations and photos in a personal digital scrapbook.
🏆 Achievements and Streaks: Track outdoor activities, maintain streaks, and unlock achievement badges.
📶 Offline-First Experience: Access saved observations and complete nature missions without an internet connection.
🔒 Privacy-Focused: Keep observations and photos stored locally in the browser instead of relying on cloud storage.

The idea is simple: use AI to get people outside, not keep them online.

WildBuddy is designed for nature lovers, students, families, and anyone looking for a healthier balance between technology and the outdoors.

🎬 Demo

The repository includes setup instructions and a demo guide.

A live deployment or video demonstration is not yet linked here.

💻 Code

Explore the complete source code here:
The project is organized into a React-based frontend, a Node.js and Express backend, automated tests, and documentation.

The repository is open source under the MIT License.

🛠️ How I Built It

I built WildBuddy using modern web technologies and open-weight AI models to make nature exploration accessible, private, and resilient even without an internet connection.

Technology stack:

Frontend: React, TypeScript, Vite, and Tailwind CSS
Backend: Node.js, Express, and TypeScript
Local AI: Ollama with open-weight vision-language models such as Gemma 3
Local Storage: IndexedDB for observations, photos, and progress
Offline Support: Progressive Web App (PWA) capabilities
Testing: Vitest for testing streak calculations, achievement badges, and nature mission generation

Ollama allows the application to run supported AI models locally rather than depending on paid cloud AI APIs.

When a local AI model is unavailable, WildBuddy can operate in demonstration mode using deterministic fallback logic. This makes it easier to explore the application without configuring an AI model first.

The project also includes offline support, so users can revisit their nature journal, complete missions, and track their progress without a continuous internet connection.

🌍 Why Does Open Innovation Matter?

Open innovation made it possible to build WildBuddy around local AI, user privacy, accessibility, and freedom from cloud dependencies.

By using open-weight models through Ollama, the application can run supported AI inference locally instead of sending every identification request to a proprietary cloud service.

This approach offers several advantages:

Privacy: Nature observations and photos can remain on the user's device.
Accessibility: Users can experiment with AI without requiring a paid cloud AI subscription.
Offline Potential: Core features can continue working without an internet connection.
Transparency: Developers can explore, modify, and extend the source code.
Community Collaboration: Other developers can contribute improvements, add features, and adapt the project for different environments.

Open innovation is not just about making code available. It is about giving people the freedom to understand, improve, and build upon technology.

With WildBuddy, I wanted to explore how open AI could encourage healthier digital habits while helping people discover the natural world around them.

🤖 My Agent Session

I used development tools to build and organize the project.

Agent session: Not linked yet.

If I publish a DevRelay session for this project, I'll add it here.

🏆 Prize Categories

Challenge: Hacktoberfest Open-Source AI Challenge — Week 1: Touch Grass

My project focuses on:

Open-source AI
Local inference with open-weight models
Offline-first applications
Nature exploration and reduced screen time

💚 Final Thoughts

WildBuddy started with a simple question:

What if AI encouraged us to spend less time using technology and more time experiencing the world around us?

Instead of creating another app that demands our attention, I wanted to build something that helps us put our phones away, step outside, and notice the nature around us.

Because sometimes, the best use of technology is helping us disconnect from it.

🌿 Less scrolling. More exploring.

Thanks for checking out my project!