Inspiration
Toyota Matchmaker was created to alleviate any apprehension a prospective buyer may have about the extensive Toyota lineup. We interviewed prospective buyers who were overwhelmed by trims, financing, and incentive programs, and aimed to provide guidance. In addition, the project serves as a legacy of Toyota Financial Services' focus on responsible mobility and safety, helping buyers feel confident in their purchase, championing the Toyota brand, and creating additional revenue-generating opportunities.
What it does
Toyota Matchmaker is a chat-based web experience that helps buyers find the best vehicles, trims, and financing options for their unique situations. For instance, buyers enter information about their lifestyle, budget, and requirements, and the assistant gathers the best offers for review. The chatbot champions the best incentives for each buyer, not only by explaining why those vehicles would work best, but also by providing transparent financing counsel. In sum, all suggestions center on a customized, well-informed, and cost-effective approach to purchasing a Toyota.
How we built it
The application runs on a Vite React and TypeScript front-end, styled with Tailwind CSS and shadcn/ui for consistent, reusable components. We sketched out the experience on paper before implementing a component-driven approach. Supabase creates a real-time persistence layer for leads and conversational context, allowing the assistant to remember previous conversations. We built a custom prompt layering system to ensure the chatbot knows only about Toyota, providing accurate, branded responses.
Challenges we ran into
Crafting the chatbot demanded careful prompt tuning to keep conversations relevant, and we lost time wrangling Gemini’s outdated API responses before falling back to a more stable interface. Integrating those fixes without breaking the UI required plenty of debugging breaks.
Accomplishments that we are proud of
We provided a smooth experience that made finding a car feel natural rather than scary. Conversations are now in line with Toyota's product catalog and brand tone thanks to the prompt orchestration layer. Every page, pop-up, and chat interaction seems consistent across devices thanks to our design system.
What we learned
We found that the correct information can create a good dialogic interface to support you through a multi-step purchase. We learned that you must maintain personal boundaries, even with an assistance tool, to ensure that any answer given is based solely on the information provided. We learned some new client state handling for Supabase, as this conversation could take hours/days. We understood the distinction between a new hypothesis/a new path of inquiry and making an impulse decision to back out of faulty API sources that would otherwise extend the time.
What is next for the Toyota Website
The next thing we will implement is dealership inventory feeds to surface real-time availability and local incentives. We want to ensure that the bilingual coverage is implemented, as many more community members who want Toyotas may benefit from our assistance. We will implement expanded analytics dashboards to give the marketing team insight into what recommendations are most compelling to shoppers. We are also in discussions with Toyota Financial Services to facilitate customer pre-qualifications to make the conversation-to-contract transition easier.
Built With
- flask
- gemini
- node.js
- react
- tailwind
- typescript
- vite

Log in or sign up for Devpost to join the conversation.