CraftCopy: Build Product Listings with Open-Source AI — A Side Hustle for Small Businesses

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

What I Built

I built CraftCopy — Local AI Listing Studio, a browser-based AI assistant that helps small businesses create product listings quickly and easily.

Many small business owners, handmade-product sellers, Instagram shops, and online resellers spend valuable time writing product titles, descriptions, and search tags. Hiring a copywriter or paying for a hosted AI service may not be practical for everyone.

CraftCopy aims to solve this problem by generating editable product listing content directly in the browser.

Users can enter their product name, category, features, target audience, preferred brand voice, and keywords. The application then generates a product title, description, and relevant search tags.

Key features

• AI-powered product title generation
• Product description writing
• Search tag suggestions
• Customizable brand tone and target audience
• Editable results with copy buttons
• Export listings as a text file
• Browser-based inference without a paid AI API key

My goal is to make AI-assisted product writing more accessible and affordable for independent sellers.

Demo

Try CraftCopy here:

Live Demo — CraftCopy Local AI Listing Studio

Enter a product name and its features, choose your preferred tone, and generate a draft listing.

Code

The project is open for everyone to explore and improve.

GitHub Repository — CraftCopy Local AI Listing Studio

The project is built with:

• HTML — page structure and product input form
• CSS — responsive layout and styling
• JavaScript — application logic and AI integration
• Transformers.js — running the model in the browser
• Xenova/flan-t5-small — the open-weight text-generation model used by the application

How I Built It

I built CraftCopy using HTML, CSS, and JavaScript, keeping the application lightweight and accessible through a web browser.

For the AI component, I used Transformers.js with the browser-compatible Xenova/flan-t5-small model. Instead of sending product information to a hosted text-generation API, the application loads the model into the browser and performs text generation on the user's device.

The application takes the user's product details and creates prompts for three tasks:

1. Generate a concise product title.
2. Write a product description based on the supplied information.
3. Suggest relevant search tags.

The results appear in an editable interface, allowing users to review, modify, copy, and export their listing.

I also implemented a fallback template so users can still receive an editable draft if the AI model fails to load. The application identifies this as template mode rather than presenting the fallback as AI-generated content.

The model and supporting JavaScript library need to be downloaded on first use, so an internet connection is initially required. Subsequent caching and offline availability depend on the browser and its stored resources.

Why Does Open Innovation Matter?

Open innovation is important for CraftCopy because the project aims to make useful AI tools more affordable, adaptable, and privacy-conscious for small businesses.

1. Privacy and control

The application performs text generation in the browser rather than sending product details to a hosted text-generation API. This gives sellers more control over their product information.

2. Lower operating costs

There is no per-request charge from a paid AI inference API. This can make experimentation and frequent listing generation more affordable, although users still need a compatible device and may incur internet and electricity costs.

3. Freedom to customize

Developers can modify prompts, experiment with compatible models, and add specialized instructions for different product categories. For example, a jewellery seller could use different listing rules from a vintage clothing reseller.

4. Less vendor lock-in

The project does not depend on a single paid AI provider's API pricing or availability. The model and application can be changed as the project evolves.

5. Learning through open technology

Using open-source tooling and an inspectable codebase makes it easier for other developers to learn from the project, suggest improvements, and contribute new features.

For this project, open innovation makes it possible to explore a small-business AI product without immediately taking on the cost and complexity of a paid inference service.

My First User

CraftCopy is designed for independent sellers, including handmade-product creators, Instagram shops, and small e-commerce businesses.

My next step is to share the application with a few sellers and ask them to create real listings with it. I want to understand whether the generated content saves them time, what improvements they need, and whether they would pay for a more advanced version.

I have not included claims about paying customers or user feedback because those should come from actual testing. I plan to use the feedback to prioritize features such as batch listing generation, saved brand profiles, and CSV export.

Prize Categories

Primary category: Side Hustle — building an AI-powered tool intended to solve a practical business problem and potentially become a paid service.

I am focusing on the challenge's open-source AI theme by using an open-weight model and browser-based inference.

---

CraftCopy is my attempt to turn a practical problem into a small, useful AI product. I hope to keep improving it through real-world testing, open collaboration, and feedback from independent sellers.

Thanks for checking out my project!