Show HN: Chemistry AI – A step-by-step chemistry solver for students (chemistryai.chat)

Chemistry AI is a Show HN project that provides a step‑by‑step solver tuned specifically for real classroom chemistry problems — not simplified toy examples. It tackles the kinds of tasks students actually encounter: word problems, reaction mechanisms, equilibrium calculations, and multi‑step stoichiometry. The tool prioritizes explicit logical steps and worked solutions rather than shortcut answers, so users can follow the reasoning and check intermediate calculations. For the AI/ML community this is a useful example of domain‑focused tuning and pedagogy‑aware interface design: rather than optimizing only for final accuracy, the system emphasizes transparent chain‑of‑reasoning to support learning and auditability. That has implications for model training and prompting strategies (favoring stepwise outputs, unit handling, mechanism detail) and for reducing silent errors common in black‑box LLM answers. It’s also notable from an education standpoint — it can serve as a tutor, homework checker, or instructor aide — while raising familiar questions about assessment integrity and how such models should be integrated into curricula.