Toyota Redline Scout

Search results and matching algorithm
Talk with Scout as you would any friend
Redline offers financing estimations
Locates and find your nearest dealership
Has a comparison feature for users to compare cars side by side
Offers AI powered insights to highlight car details

Inspiration

Have you ever tried to buy a car only to be met with difficult customer service and a bunch of technical jargon that you didn’t understand? Next, the dealer salesman tried to rip you off with inflated prices and untrustworthy sales tactics. The entire process is low, frustrating, and inefficient….

Until now. Meet Redline, the all inclusive AI-powered research and automation tool that expedites your journey to your new automobile. Using state of the art algorithmic LLM processing and web scrapping technology, Redline automatically researches automobiles for you based on your exact preferences. With semantic search and inventory awareness, Redline is the new best way to buy a car.

Talk to the AI assistant as you would talk to a friend. Use natural English without the need for technical terms. Redline will then take your human responses and utilize the leading models from OpenAI to match you with recommendations that perfectly suit your needs.

Want to adjust your preferences? No worries, Redline will adapt to your changes in real time. Looking to buy? Redline offers a multitude of financial estimation features, based on down payment, loan term, credit score, and max monthly payment.

Redline is revolutionizing the consumer experience in the automobile industry: skip the line, avoid the customer service BS, use AI to optimize, and buy online.

What it does

Ask user for location for dealership search
Offers a chatbot feature that allows users to interact with it as a normal human
The matching algorithm takes care of all the technicals, reduces ambiguity and complexity
Based on the conversation, the AI assistant will use NLP to extract parameters that fit the desired preferences of the user
Uses AI matching algorithm to related parameters to keywords within the web-wrapped data set to create custom car recommendations
Asks about financing information for a financing estimate
Interacts with user to match their wants and needs to dealer inventory, adjusts to changes in real time
Shows cars and their info, user can choose multiple cars at a time to compare and contrast
Highlights a brief AI overview of each recommendation and why it was recommended

How we built it

Backed built mostly with Python and Node.js, frontend with React
Web scraped data using Python libraries such as Beautiful Soup to create JSON file database containing detailed vehicle info
User’s natural language input is fed to an OpenAI client, that extracts keywords and matches it to backend search parameters
Implements Openrouter to connect multiple AI models
- Uses OpenAI's o4-mini to quickly process user frontend responses for a smooth UI/UX
- Uses GPT-5.2 to algorithmically match labels to keywords in our database to create recommendations for better accuracy
After each user input/interaction, the user preferences are updated
Show cars uses matches user preferences to listings from the database

Challenges we ran into

Wrote a custom data gathering script to use real-life listings and accurate data
Markdown errors using native React API, had to write custom markdown renderer
Working with OpenRouter keys to utilize two models, each which were more optimized for their task