Inspiration
Queens is the most linguistically diverse place on Earth, yet newly arrived immigrants often navigate legal, health, and social systems entirely alone. We wanted to build the first resource a newcomer reaches for, something that feels like a knowledgeable friend, not a government portal.
What it does
One Queens lets users search for legal aid, ESL programs, healthcare, food assistance, housing, and employment resources in Queens which is filtered by neighborhood, language, and cost. A smart assistant answers plain-language questions in any language, pulling live web results when needed. Ask "Where can I find a free immigration lawyer in Jackson Heights?" and get a specific, cited answer with real addresses and phone numbers.
How we built it
React + Vite frontend with static resource pages per category. Python Flask backend with a tool-calling loop: the model decides whether to search the web via SearXNG, retrieves results, and synthesizes a cited response.
Challenges we ran into
Tuning the tool-calling loop so the model searched only when necessary and didn't loop indefinitely. Getting the tone right for users who may be stressed, unfamiliar with US systems, and writing in languages other than English.
Accomplishments that we're proud of
Multilingual support, real organization data with actual addresses and phone numbers, a "Know Your Rights" section for ICE encounters and tenant rights, and a clean mobile-first UI built in 12 hours.
What we learned
How to build a reliable agentic tool-calling loop, how much grounding AI in local data matters over generic answers, and how to scope ruthlessly for a hackathon.
What's next for One Queens
A step-by-step "I Just Arrived" onboarding checklist, expansion to all five boroughs, offline PWA support, and full translation of resource pages into Spanish, Mandarin, Bengali, and Urdu.
Log in or sign up for Devpost to join the conversation.