Mint Rentals

Inspiration

Mint Rentals came from a problem we have all experienced as college students: finding housing is stressful, time consuming, and overwhelming. Students have to balance rent prices, commute times, neighborhood safety, and housing availability, all while juggling classes and work. Most platforms only solve one piece of the problem, forcing users to switch between multiple apps and websites to make a decision. We wanted to simplify that experience into one streamlined platform that helps students quickly find rentals that actually fit their needs.

What it does

Mint Rentals is an AI-powered rental search assistant built specifically for students. Users enter preferences such as budget, preferred streets or neighborhoods, square footage, and other housing requirements. Our agent then scrapes platforms like Craigslist and Redfin to find available rental listings that match those preferences.

Beyond just listing rentals, Mint Rentals also evaluates practical factors that matter to students. The platform analyzes nearby bus routes and commute accessibility to the university, while also using national crime and safety datasets to rank neighborhood safety. By combining pricing, location, transportation, and safety into one workflow, Mint Rentals helps students make faster and more informed housing decisions.

How we built it

We built Mint Rentals using a React and TypeScript frontend with a clean UI prototyped in Figma. For the AI workflow, we used NVIDIA DGX Spark hardware connected through a laptop setup, running models through Ollama and NVIDIA infrastructure. We integrated OpenClaw as part of the agent pipeline to process user preferences and automate rental discovery workflows.

The system combines web scraping, route analysis, and safety data aggregation into a single pipeline that delivers curated rental recommendations to users in real time.

Challenges we ran into

One of the biggest challenges was setting up and connecting the DGX Spark hardware correctly. Many teams experienced issues with device naming and connection configurations, which made the setup process surprisingly difficult during the hackathon.

Another major challenge was combining data from multiple sources into a cohesive experience. Rental platforms, transportation data, and neighborhood safety datasets all use different formats and structures, so normalizing and organizing that information in a reliable way took significant effort. We also had to optimize scraping and response times so the system could return useful recommendations quickly enough for a live demo.

Accomplishments that we're proud of

We are proud that we built a working end-to-end platform instead of just a simple housing search prototype. Mint Rentals not only finds listings, but also adds meaningful context around transportation and neighborhood safety that students actually care about when making housing decisions.

We are also proud of successfully integrating AI workflows with DGX hardware acceleration during the hackathon. Despite infrastructure and setup challenges, we were able to create a functional system that ties together scraping, ranking, and recommendation features into a polished user experience.

What we learned

We learned that building reliable AI systems is not just about the model itself, but about integrating data pipelines, infrastructure, and user experience into one coherent workflow. We also gained hands-on experience working with NVIDIA DGX hardware, Ollama deployments, and AI agent orchestration under hackathon time constraints.

Another important lesson was how valuable contextual information is in decision-making. Students do not just want a list of apartments — they want confidence that the location is affordable, safe, and accessible to campus.

What's next for Mint Rentals

We want to expand Mint Rentals beyond student housing and make it a more comprehensive rental intelligence platform. Future plans include adding support for more rental marketplaces, improving commute analysis with real-time transit data, and incorporating personalized recommendation systems based on lifestyle preferences.

We also want to build stronger filtering, interactive maps, roommate matching, and predictive pricing insights so users can better understand both the housing market and the tradeoffs between cost, safety, and convenience.

Built With

• dgx-spark-/-asus-ascent-gx10
• fastapi
• javascript
• nemoclaw
• nvidia-nemotron
• openclaw
• python
• react
• tailwind-css
• vercel
• vite
• vllm