CraftCopy AI: Open-Source AI Copywriting for Small Businesses

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

What I Built

CraftCopy AI — Your products, better words.

CraftCopy AI is a browser-based AI writing assistant built for small business owners, handmade product sellers, freelancers, and independent creators.

Writing product listings and social media captions for every product can be time-consuming, especially for people running a small business alongside college, work, or other responsibilities.

CraftCopy AI helps simplify this process. Users enter their product name, category, details, target audience, preferred brand voice, and selling platform. The application generates marketing copy tailored to their inputs.

Key features

• Generate product descriptions, titles, social media captions, and hashtags.
• Choose a brand voice, from cute and playful to elegant and professional.
• Create copy for platforms such as Instagram, Etsy, Shopify, and WhatsApp Business.
• Copy generated content for editing and publishing.
• Run open-weight AI inference directly in the browser without requiring a paid AI API key.

The goal is simple: help small sellers spend less time writing listings and more time building their businesses.

Demo

🚀 Try CraftCopy AI: https://narbiya-132.github.io/CraftCopyAI/

The project is hosted on GitHub Pages. The AI model may need to download when used for the first time, so an internet connection is required for initial setup.

Code

💻 GitHub Repository: https://github.com/narbiya-132/CraftCopyAI

The project uses HTML, CSS, and JavaScript, with Transformers.js for browser-based model inference.

How I Built It

I built CraftCopy AI using three core files:

• index.html — the user interface and product input form.
• style.css — the responsive layout, styling, and visual design.
• app.js — form handling, model loading, prompt construction, text generation, and copy-to-clipboard functionality.

The AI component uses Transformers.js with the Xenova/flan-t5-small model, an open-weight text-generation model.

Instead of sending every product description to a proprietary AI API, the application loads the model in the browser and performs inference locally.

The workflow is straightforward:

1. The user enters product details and selects their preferred output.
2. JavaScript constructs a prompt using those details and preferences.
3. Transformers.js loads the open-weight model.
4. The model generates marketing copy in the browser.
5. The user reviews the result and copies it for further editing.

This approach avoids a dependency on a paid, hosted inference API. However, model downloads, browser compatibility, and available device resources can affect the experience.

Why Does Open Innovation Matter?

Small businesses often have limited budgets, and recurring AI API costs can become another barrier to getting started.

Open-weight models offer an alternative: developers can build applications around models they can run, evaluate, and potentially replace without being tied to one proprietary inference provider.

For CraftCopy AI, open innovation matters in three ways:

1. Lower barriers to entry

A seller-focused tool should be affordable to operate. Browser-based inference avoids per-request charges from a paid AI API, although hosting, development, and other infrastructure can still have costs.

2. Greater control over data

Product details can be processed locally by the model after it has loaded. This reduces the need to send those details to a third-party AI inference server. The application still relies on external resources for its initial model download, and browser-based processing is not the same as complete offline operation.

3. Freedom to experiment

Using an open model and an open-source inference library makes it easier to experiment with prompts, compare models, and adapt the application for specific niches, such as handmade gifts, jewellery, stationery, and small online shops.

A closed API can offer stronger results or more advanced capabilities in some cases. The advantage here is the opportunity to build a practical, adaptable tool without making a paid proprietary API the core of the application.

For me, this project is a step toward making useful AI tools more accessible to people starting small businesses.

Prize Categories

Open-weight AI / open-source AI application.

I'll confirm the exact partner category names against the challenge rules before final submission.