BoomCheck AI

Decode Danger. Protect Health. One Upload Away.

Comment

Inspiration:

BoomCheck AI was inspired by a simple but critical observation: many students and healthcare professionals work with chemicals in labs without fully understanding the associated risks. Whether in academic settings or healthcare environments, exposure to unknown chemical hazards can lead to serious safety issues. In chemistry courses, lab procedures are often buried in dense documents, and Material Safety Data Sheets (MSDS) are either ignored or too time-consuming to interpret. We wanted to automate this process and provide instant, AI-driven safety guidance from any lab document.

What it does:

BoomCheck AI extracts chemicals from uploaded lab documents, summarizes MSDS data, predicts hazards when chemicals are combined, and offers AI-generated safety tips, safer alternatives, quizzes, and lab procedures—all in a sleek, drag-and-drop web interface.

How we built it:

  1. Frontend: React.js + Tailwind CSS for a chemistry-themed, drag-and-drop UI
  2. Backend: Node.js + Express with file parsing (PDF/DOCX) and OpenAI/ GeminiAI integration
  3. AI Module: Uses GeminiAI/OpenAI to extract chemical names, predict reactions, advise safety, suggest safer alternatives, generate pre-lab quiz and summary of lab procedure
  4. Database: MongoDB stores MSDS for known chemicals
  5. Routing: RESTful API routes handle uploads, chemical info, combination analysis, quizzes, and procedure generation

Challenges we ran into:

  1. Reducing latency from multiple AI calls (we solved this by batching requests and caching results)
  2. Designing an intuitive UI that still communicates complex scientific information
  3. Balancing AI predictions with rule-based safety logic for more reliable outputs
  4. Extracting accurate chemical names from inconsistent document formats
  5. Scale our database

Accomplishments that we're proud of:

  1. A fully working end-to-end chemistry safety assistant in under 24 hours
  2. Real-time hazard analysis with AI-powered alternatives and safety messages
  3. Smooth drag-and-drop UX tailored for lab learning and experimentation
  4. Seamless integration of NLP, MSDS summaries, and AI content generation
  5. Dynamically background

What we learned:

  1. How to combine NLP, AI, and rule-based logic into a single backend system
  2. Tailwind and React best practices for quick front-end iteration
  3. Prompt engineering for reliable and fast OpenAI API responses
  4. How to design for both educational and professional healthcare audiences
  5. Choice of database, structure
  6. How to improve performance of the frontend and backend

What's next for BoomCheck AI:

  1. Replace the MongoDB with live chemical info from PubChem or ChemSpider
  2. Fine-tune AI to improve chemical name recognition and hazard prediction accuracy
  3. Add OCR support for handwritten/scanned lab sheets
  4. Better improve performance
  5. Expand support to hospital and pharmacy labs for real-world healthcare impact

Built With

Share this project:

Updates