RentGuard

Inspiration

The rise of online rental platforms has created both opportunities and risks. While millions of people use sites like OLX, Facebook Marketplace, and WhatsApp groups to find affordable housing, the prevalence of fraudulent listings has led to significant financial losses and emotional stress. Fake landlords, duplicate images, and deceptive conversations continue to trap unsuspecting tenants. Our team was inspired to build RentGuard after seeing how common these scams are in our own communities. We wanted to create a solution that leverages the power of Agentic AI to protect renters and restore trust in the rental ecosystem.

What it does

RentGuard is an AI-powered scam detection and prevention system designed specifically for the rental market. It analyzes both the text and images of rental listings to identify inconsistencies and red flags. It further performs reverse image searches to detect reused or stolen photos. Beyond passive detection, RentGuard actively engages with suspected scammers using AI-driven conversations to uncover fraudulent behavior. In the end, users receive a clear, simple outcome: whether the listing is trustworthy or potentially fraudulent.

How we built it

We built RentGuard using a combination of LLMs (Large Language Models) for text verification and computer vision models for image analysis. The backend was developed using Python and Node.js, while APIs were integrated for reverse image checks and conversational automation. A lightweight browser extension was designed to make RentGuard easily accessible on popular platforms where rental ads appear. To enhance performance, we curated a dataset of genuine and fraudulent listings to refine the detection pipeline. This hybrid approach of text, image, and interactive engagement forms the backbone of RentGuard’s unique verification model.

Challenges we ran into

One of the biggest challenges was balancing speed with accuracy, as real-time detection is critical for user adoption. Creating convincing AI-driven conversations that expose scammers without revealing the system’s identity also required extensive testing. Integrating multiple services—LLMs, reverse image APIs, and web scraping tools—into a seamless workflow posed technical difficulties. Additionally, the scarcity of publicly available labeled scam data made it necessary for us to manually curate and validate our dataset. Despite these hurdles, we successfully developed a functional prototype that addresses the problem effectively.

Accomplishments that we're proud of

We are proud that within the hackathon timeframe, we created a working prototype that demonstrates measurable impact. Our system successfully detected more than 70% of fraudulent listings in test scenarios. A unique achievement was the design of an AI-to-scammer interaction mechanism, which is rarely attempted in conventional fraud detection solutions. Furthermore, RentGuard’s browser extension provides a user-friendly way for renters to protect themselves without needing any technical expertise. Most importantly, we are proud of building a project with the potential to evolve into a real-world product that can make renting safer.

What we learned

Through this project, we learned the immense potential of Agentic AI in solving practical, real-life problems. We discovered that effective fraud detection requires a multi-modal approach that combines text, image, and conversational analysis. We also gained experience in dataset creation, pipeline integration, and API orchestration under tight time constraints. More importantly, we realized that building trust-based technology is not only about technical accuracy but also about designing a simple, reliable user experience that inspires confidence.

What's next for RentGuard

Moving forward, we plan to develop a mobile-first version of RentGuard that can directly scan listings shared via WhatsApp and Telegram, where scams are most prevalent. We aim to expand our training dataset to further increase accuracy and reduce false positives. Collaborations with housing platforms and rental services will help integrate RentGuard as a built-in safety layer for users. In the long term, our vision is to scale RentGuard into a global platform, extending beyond rental properties to other domains of digital marketplaces where fraud is a growing concern.

Built With

• fastapi
• github
• javascript
• langchain
• postgresql
• python
• sqlite
• tailwindcss