OpenWorldAI

• 

  Smart search functionality to help users find AI tools they can use.

Inspiration

During the beginning of 2023, we were having a difficult time keeping up with all the cool new projects and developments in the AI space. Additionally, there was also a difficulty with understanding which tools we could implement into our workflows and daily lives. Ease of discover and ease of use are the biggest barriers to AI adaptation at this moment, in our opinion. So our initial project was to create a resource website with a UX that will help users discover new projects and understand how each project can help users. We used OpenAI to classify the projects and categorize them. For this hackathon, we have added semantic search to help users understand what projects they can start using daily. Many use-cases get lost in databases and directories because of inefficient search functionality.

What it does

We have a collection of over a thousand AI projects and companies that we classified, wrote detailed descriptions of what they do through scrapping the project html, parsing it and feeding it into OpenAI. Then all of that data was fed into the Pinecone database.

This data is available through a user friendly search on the website that helps to identify what projects the user can implement into their business, work, and daily life.

How we built it

This project is built with

• Next.js and Vercel on the frontend
• OpenAI for various tasks such as classifying and content writing for each AI project and plugin
• Pinecone for Vector database for the semantic search.

Challenges we ran into

• Displaying the data that will be useful for the end consumer. To make sure the user is able to utilize the search functionality to easily find the AI tools that they can use in their daily lives.

Accomplishments that we're proud of

We are proud that we were able to finish this project in a short period of time. To implement multiple sources of information for our Pinecone database and presenting it to the user in an easy to implement way.

What we learned

• What Pinecone and other AI development tools are capable off
• Working with vector databases
• Connecting all the tools (OpenAI, Pinecone, Langchain, Next.js & Vercel) in a single project for a single purpose
• How we can better assist the user in implementing AI to make life easier and more efficient

What's next for OpenWorldAI

1. Right now we are excited to assist users with finding AI tools and solutions they need to work better and faster and to make their lives easier. But this is just the beginning.

2. Our goal for the end of the year is to become an AI "AppStore", a one stop solution where users can come to the website and implement with the solution they need with little friction. The current signup flow is tedious for users. We are looking to unify that by releasing a single sign-on/subscription functionality.

Built With

• next.js
• openai
• pinecone
• planetscale
• prisma
• tailwindcss