VoiceNest

Welcome to VoiceNest
Quickly find homes near you with our intelligent AI bot
Helpful response curated to user's request!

Inspiration

Our original idea was a similar idea with image comparison instead, as we thought it would be useful to many people, but a lot of challenges arose in that. We decided to branch off a bit and come up with something a tad bit different.

What it does

This website takes speech from the user and based on what was selected, it would find real life listings of similar houses for the user.

How we built it

The frontend is done in React, JavaScript, CSS, and HTML, and the backend is done in Python. To connect the backend with the frontend, we used the Flask API. We used three AI models that worked together, with one to analyze the user request, another to read scraped data, and the last to generate a response to the user from the collected data. We also utilized the Azure OpenAI Studio and Apify to scrape Zillow information.

Challenges we ran into

Since we had never done it before, we had some issues implementing the text-to-speech feature. Engineering the AIs to talk/respond to each other also proved difficult because we are utilizing three models. Scraping the web and parsing the data was also something new to us, however, the biggest challenge that we faced was connecting the frontend and backend together.

Accomplishments that we're proud of

Because we implemented a lot of new ideas for this project, creating a workable demo is definitely something that we are proud of. Collaborating with multiple AI services to create a solution is also something we are delighted to implement. And although it seems small, animating the little things like the microphone button and the words on the screen make our website look that much more appealing.

What we learned

We learned a lot about react and how things worked to make the front end look nicer. Getting AI models to interact with each other was also interesting. This was the first time any of us had utilized scraping, which was exciting.

What's next for VoiceNest

We hope to improve VoiceNest by expanding types of user input (drawings, text) and expanding locations so that the website will work in a greater area.