Homestead

Inspiration

Homestead was inspired by a simple but frustrating reality: homeownership is still full of hidden rules, scattered programs, and confusing paperwork, especially for first-time and lower-income buyers. We wanted to build something that helps people answer the question, “Can I actually afford this home, and what help am I missing?” instead of forcing them to guess. The idea grew from the belief that housing should feel navigable, not intimidating.

What it does

Homestead is an AI-powered housing decision-support app that helps users understand affordability and eligibility for housing assistance programs. It takes a buyer’s profile and a property listing, then ranks homes by fit, estimates monthly costs, and shows which programs they may qualify for, such as FHA, VA, USDA, or HomeReady. It also explains why a user may not qualify and links them to official HUD and CFPB resources so they can take the next step with confidence.

How we built it

We built Homestead with a React + Vite frontend and a FastAPI backend for the eligibility engine. The system uses explicit, rules-based logic rather than a black-box model, so every recommendation is auditable and easy to explain. We combined user profile inputs, property details, and public program rules to generate affordability estimates, eligibility results, and lender-rate rankings.

Challenges we ran into

One major challenge was handling how fragmented housing rules are across federal, state, and local programs. Another challenge was designing the app so it could be helpful without pretending to make final financial or legal decisions. We also had to balance realism and demo usability by working with curated listings, synthetic examples, and public guideline-based logic.

Accomplishments that we're proud of

We’re proud that Homestead does more than just calculate numbers — it turns complex housing data into understandable guidance. The app clearly shows both qualifying and non-qualifying results, which makes the experience more transparent and trustworthy. We also built a system that keeps humans in the loop, reinforcing that lenders and HUD-approved counselors make the final decisions.

What we learned

We learned that the hardest part of building useful AI is often not the model itself, but the structure around it: clear inputs, transparent rules, and responsible output. We also learned how important it is to design for trust when the topic affects someone’s finances and future. Most of all, we learned that good decision-support tools should reduce confusion, not add to it.

What's next for Homestead

Next, we’d like to expand Homestead with more local assistance programs, better lender comparisons, and richer affordability modeling. We also want to improve the listing experience with live data integrations and more personalized guidance for different buyer types. Over time, Homestead could become a broader housing navigator that helps more people move from uncertainty to action.

Built With

• apis
• browser-localstorage
• docker-compose
• javascript-(react-19-+-vite)
• npm-workspaces
• picsum
• postgresql-with-postgis/pgvector
• pydantic
• python-(fastapi
• redis
• simplyrets
• sqlalchemy/alembic
• uvicorn)