Chemistry AI

Inspiration

I kept seeing the same pattern: students take a photo of a chemistry problem, get a short answer somewhere, and still don’t understand why it’s correct. I wanted a tool that teaches the reasoning like a patient tutor—especially for exam-style questions.

What it does

Chemistry AI solves chemistry problems from a photo or typed text and explains the solution step by step. It helps with:

• Stoichiometry & limiting reagents
• Acids & bases, pH, buffers
• Thermochemistry, kinetics, equilibrium
• Redox and common reaction patterns It also tries to call out common mistakes (units, setup errors, missing steps) so the result is actually learnable.

How I built it

The app is a web-first workflow: upload a screenshot or paste a question → parse the problem → generate a structured solution → render it cleanly with math formatting. I focused on clarity, not just correctness, and optimized prompts to show the setup and reasoning before the final answer.

Challenges

The hardest part was making explanations consistent across different question styles (word problems, multi-step reactions, unit-heavy calculations). Another challenge was handling messy screenshots and ensuring the output stays readable.

What’s next

More topic coverage, better diagram support, and stronger “check my work” feedback for students.

Project link: https://chemistryai.chat

Built With

• docker
• latex/katex
• llm-api
• next.js
• node.js
• ocr-(image-to-text)
• react
• tailwind-css
• typescript