Side_Hustle

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

What I Built

I built ListingLift AI, a web-based product listing assistant designed for small business owners, freelancers, and online sellers who want to spend less time writing product descriptions.

Selling a product online involves more than just uploading a photo. Sellers often need to write a clear title, explain the product's features, identify their target audience, and choose relevant search tags. Doing this repeatedly for multiple products can become tedious.

ListingLift AI aims to simplify that process. Users enter a product name, product details, target customer, marketplace, and relevant keywords. The app then prepares a listing title, product description, and search tags that the seller can review and copy.

I built this project around a simple idea: useful AI tools should be accessible to small sellers, not just businesses with large budgets.

The potential users are independent sellers and small businesses that need help with repetitive product writing. A future version could offer additional marketplace-specific templates and business features. I have not yet validated whether customers would pay for it, so that remains an important next step.

Demo

Try ListingLift AI:

https://vidish0008-prog.github.io/Side_Hustle/

The website is deployed using GitHub Pages.

The interface includes a demo fallback so visitors can explore the workflow without a local AI model. Actual local AI generation requires Ollama and a compatible model to be installed and configured on the user's computer.

Code

GitHub repository:

https://github.com/Vidish0008-prog/Side_Hustle

The project uses three main files:

• index.html — the website structure and input form.
• style.css — the responsive interface and visual design.
• script.js — the listing-generation workflow, Ollama integration, demo fallback, and copy-to-clipboard functionality.

The repository also includes a README with setup instructions and notes about the current limitations.

How I Built It

I built ListingLift AI using HTML, CSS, and JavaScript. I wanted the project to be approachable as a beginner project while still exploring a practical application of open-source AI.

The application is designed to connect to Ollama, a tool for running language models locally. Its AI integration requests the qwen2.5:3b model through Ollama's local HTTP API.

The workflow is straightforward:

1. The user enters the product details and selects the marketplace and writing style.
2. JavaScript prepares a prompt containing the supplied product information.
3. When Ollama is available and properly configured, the prompt is sent to the local model.
4. The model is asked to return a structured response containing a title, description, and tags.
5. The user can review and copy the generated content.

If the local model cannot be reached, the app displays a clearly labelled, template-based demo preview instead. That distinction matters because a template-based preview is not the same as AI-generated content.

The project is currently a front-end prototype. It still needs further testing, improvements to local model setup, and feedback from actual sellers.

Why Does Open Innovation Matter?

For this project, open innovation is about giving small businesses more control over their tools and costs.

1. More control over costs

A locally running model avoids a paid, hosted-model API for each generation request. This could make the product more affordable for small sellers who need to generate listings regularly. Local computing still has hardware and electricity costs, and the model needs to be downloaded first.

2. Better control over data

When the app runs with Ollama locally, product details can be processed on the user's own computer instead of being sent to a third-party hosted AI provider. This is useful for sellers who prefer to keep their business information under their control.

3. Less dependence on a single provider

A local-model approach gives the project room to experiment with different models as the application evolves. It also creates opportunities to test models for different writing styles, product categories, and marketplace requirements.

4. Learning by building

Open tools made it possible for me to experiment with a practical AI workflow without starting with a paid AI API. I could build the interface, connect it to a local model, and create a fallback experience when the model is unavailable.

There are trade-offs. Local inference depends on the user's hardware, setup can require extra configuration, and the output still needs human review.

I see open innovation as more than making a tool free. It is about giving people more freedom to understand, adapt, and control the technology they use.

What I Learned

This project helped me explore how a regular website can interact with a locally running AI model. I also learned that a good AI product needs more than an attractive interface: it needs a clear problem, a reliable generation workflow, transparent failure handling, and feedback from real users.

My next goal is to test ListingLift AI with a small online seller, learn which parts of the listing process take the most time, and use that feedback to improve the project.