SafeHouse is an AI-powered web application that helps users evaluate the safety, health, and livability of their homes using image-based analysis. It detects environmental hazards, explains risks in simple language, and provides actionable next steps like repairs, services, and essential products.
Most people move into or live in spaces without understanding hidden risks such as:
- Mold and moisture damage
- Water leaks
- Unsafe electrical wiring
- Structural hazards
- Poor living conditions affecting mental and physical health
Traditional inspections are expensive, slow, and not accessible to everyone.
SafeHouse turns any living space into a real-time AI safety check system.
Users can upload images of their home, and SafeHouse will:
- Detect potential hazards using AI vision
- Explain issues in simple, non-technical language
- Recommend specific repairs needed
- Suggest nearby repair shops (with best-rated options)
- Provide essential product recommendations (e.g., Amazon items)
- Highlight accessibility and neurodivergent-friendly concerns
- 🖼️ AI-based hazard detection from images
- 🛠️ Repair recommendations for identified issues
- 📍 Nearby repair shop discovery + ranking
- 🛒 Essential home product suggestions
- 🧠 Simple explanations of risks and fixes
- ♿ Neurodivergent-friendly structured outputs
- 🏠 Move-in safety insights for renters & students
- User uploads an image of a room or space
- Computer vision model detects possible hazards
- LLM layer translates detections into human-readable insights
- System generates:
- Risk summary
- Repair actions
- Local service recommendations
- Essential product suggestions
- Results are displayed in a clean, structured UI
Python, JavaScript, React, Next.js, Flask, FastAPI, OpenAI API, Computer Vision, OpenCV, REST APIs, Geo-location API, Web scraping, Amazon API, HTML, CSS, TailwindCSS
- Converting visual detections into meaningful real-world advice
- Mapping AI outputs to actionable repair categories
- Ranking hazards by severity in a useful way
- Integrating external services (shops + products)
- Keeping results fast and user-friendly
- Built an end-to-end AI-powered safety system
- Combined computer vision with LLM reasoning
- Designed actionable outputs (not just detection)
- Integrated real-world repair and product ecosystems
- Focused on accessibility and cognitive clarity
- How to combine vision models with language models effectively
- Designing AI for real-world decision support
- Importance of accessibility in AI systems
- Turning raw predictions into human impact
- 📱 Real-time camera scanning mode
- 🧠 Improved hazard detection accuracy
- 🗺️ Smarter geo-aware repair recommendations
- 🏠 “Home Safety Score” for rentals and apartments
- 🤝 Integration with rental platforms
- 🧑⚕️ Environmental health risk predictions
- 💻 GitHub Repository: https://github.com/shreya222mishra/safeHousing
- 🌐 Devpost Submission: https://devpost.com/software/safehouse-tqx2i5
SafeHouse aims to make every home a safe, understandable, and health-aware space — turning housing from a blind risk into a guided, AI-powered decision.