Dead Stock Dost

This is a submission for the Hacktoberfest Weekend Challenge: Build for a Friend

What I Built

I built Dead Stock Dost, a tiny local-first AI tool for a small neighborhood garment shop.

A small shop can have shirts, kurtas, jackets, and other products sitting on shelves for weeks or months. Even with an inventory CSV, it isn't always easy to quickly answer:

• Which products need attention?
• Why might they be stuck?
• What could I try?

Dead Stock Dost takes a simple inventory CSV and turns it into:

CSV → Identify Dead Stock → Explain → Suggest an Action

The application deterministically identifies products that have been sitting unsold for too long, then lets the shopkeeper ask a local AI assistant for an explanation and possible actions.

The AI doesn't make the decision for them.

The code calculates the facts. The AI explains the facts. The shopkeeper decides.

Demo

🎥 Video Demo: https://youtu.be/U6-7b7IAlMg

The complete demo takes less than two minutes:

Open the app
   ↓
Try demo data / Upload CSV
   ↓
See products needing attention
   ↓
Select a product
   ↓
"Why is this stuck?"
   ↓
"What can I try?"
   ↓
Get a local AI response

Code

🔗 GitHub Repository: https://github.com/aimanfazal/dead-stock-dost

The repository contains the application, demo CSV, screenshots, architecture documentation, and project scope.

How I Built It

Dead Stock Dost is built with:

• Next.js
• React
• TypeScript
• CSV
• LM Studio
• Qwen 2.5 3B Instruct

The AI runs locally through LM Studio using an open-weight model.

The application itself handles the inventory logic deterministically:

• 90+ days: Dead Stock
• 45–89 days: At Risk
• Less than 45 days: Normal
• Stock ≤ 0: Ignored

The AI receives the calculated facts for a selected product and helps explain the situation or suggest possible actions.

It does not classify products, modify inventory, change prices, invent facts, or make decisions automatically.

I intentionally kept the architecture small: no database, authentication, cloud AI API, vector database, or multi-agent framework.

The CSV remains the source of truth.

Why Does Open Innovation Matter?

Open innovation made it possible to build this as a local-first AI tool instead of depending on a paid cloud AI API.

Using an open-weight model through local inference means the shopkeeper's inventory data can stay on their own machine while still getting a natural-language AI assistant.

It also makes the project easier to experiment with and adapt without tying the core application to a single hosted AI provider.

For this project, the important part isn't making AI autonomous. It's making a small, useful AI assistant available while keeping the shopkeeper in control.