Inspiration

Finding flats and student messes in Bangladesh often means trusting unverified Facebook posts and To-Let signboards. Fake landlords, advance payment scams, stolen photos, and hidden charges make renting risky. We built RentRadar to help tenants make safer rental decisions.

What it does

RentRadar checks rental ads for stolen photos, suspicious prices, and advance payment scams. It also helps users find properties, verify landlords through utility records, calculate actual rental costs, and check lease agreements against the Premises Rent Control Act, 1991.

How we built it

We built RentRadar using Google Antigravity, Next.js 15, TypeScript, Tailwind CSS, Supabase, Gemini 2.5 Flash, Sharp, Leaflet, and OpenStreetMap. The system uses image hashing, text extraction, and rule-based checks to examine rental listings.

Challenges we ran into

Handling stylized Bangla text, staying within Vercel's upload limits, protecting sensitive documents, and getting Leaflet maps to work properly with Next.js were our main challenges.

Accomplishments that we're proud of

We built five working modules in one platform, including Bangla and English listing inspection, rental cost calculation, utility-based landlord verification, and a rent agreement assistant.

What we learned

We learned that AI alone is not enough for reliable fraud detection. Combining text extraction with direct calculations and fixed rules produces more dependable results. We also learned the importance of protecting user data.

What's next for RentRadar

We plan to expand our verified rental database, work with local utility providers and municipalities, improve rent estimates across more neighbourhoods, and make RentRadar available to more tenants across Bangladesh.

Built With

Share this project:

Updates

Submission history