Product Pal

For your tech needs.
For your clothing needs.
For your entertainment needs.

Inspiration

How often have you wanted to purchase a clothing item like a hoodie, but wanted to see similar ones on different brands and websites? Or the store that had it for the most affordable price?

Product Pal is a tool that gets you out of your online shopping frenzy. Being a proud competitor in Noibu’s track for a more delightful e-commerce experience, it is a guaranteed must for shoppers.

In order to fully get a grasp of people’s shopping habits, we asked a few hackers about how much research they do on retail items when they’re purchasing them. The general consensus was that they go to the website of a brand they trust, and buy exactly what they need. They also mentioned that they’d be open to trying out new brands given that the product information would be readily available.

What it does

Product Pal has a clean and user-friendly interface, making it easy for users to find the product they are looking for. The products are displayed with all relevant information, including images, prices, and descriptions, so users can quickly assess their options and make informed decisions. The tool also allows users to check out the products in a new tab, providing a seamless and efficient shopping experience.

How we built it

The implementation of Product Pal involved the use of several technologies, including Javascript, HTML, and CSS. These technologies were used to create a browser extension that integrates seamlessly into the user's shopping experience.

The extension utilizes Natural Language Processing (NLP) and OpenAI’s API to break down and summarize the key adjectives and nouns associated with the product the user is currently looking for. By using this information, Product Pal can provide a more accurate and relevant list of alternative products.

To display the alternative products, the team utilized the Custom Search JSON API (Programmable Search Engine). This technology allows them to programmatically search for products based on the user's query and display the first ten results in the extension. The data extracted from the search results were then used to populate the user interface component of the tool.

Challenges we ran into

Developing a Chrome extension was a challenging task, especially when it comes to the backend part. The security permission issues with Google while trying to retrieve the current browser URL is a common challenge faced by many developers. To overcome this, we made a smart decision by using Google's tabs API instead of working against them.

Additionally, summarizing the contents of the URL with Cohere's API was difficult, which led us to switch to OpenAI's API, which worked really well for us. The integration of OpenAI's API helped in summarizing the contents of the URL and made it easier for the search API to understand the content.

Lastly, finding relatable products similar to the current one on the window was a challenge that we faced. To overcome this, we used Google's custom search API, which queries links for us. However, the issue of retrieving images was encountered, and to handle that, we included a default "empty state" with the project's logo in case no image was present.

In conclusion, the challenges we faced while developing a Chrome extension are common and can be overcome with careful planning and innovative solutions.

Accomplishments that we're proud of

The accomplishment we’re the proudest of is getting this dev post ready in time!! It was already a time challenge to get a working product done by the deadline so we are really glad we got to finish this Devpost for you to enjoy and learn more about us and our project. On the technical side we were really proud to see so many of the issues we faced (refer to challenges we encountered) were able to be overcome so quickly and with a creative solution. Working with no sleep and a lot of red bull we were shocked and amazed at how well we worked together to find workarounds to issues we would individually take days to fix.

What we learned

We learned a lot of things from this hackathon project, but more specifically how to build a chrome extension and work with OpenAI. Building a chrome extension was extremely challenging but we learned the difference between when we could use a JS framework vs when we could not. When we could use ES6 or when we had to use only JavaScript for making the chrome extension, the answer boils down to static information vs dynamic information. We had to use JavaScript for the dynamic information because we had to use the OpenAI API to get the information from the chrome page itself. OpenAI's API taught us as developers how to give GPT3 prompts to use and what the best way to make these prompts are.

What's next for Product Pal

We want to scale Product Pal to not only account for retail items, but also services, including flight tickets. We understand that this is a demand many of our customers may have due to the sheer price of travelling. This is something we want to work towards making more accessible for everyone by providing the best possible prices.