spent around 3k rupees on tools and built a full pharmacy billing app with ai. now testing in a real pharmacy and scanning is giving me headache
retained text
not a proper dev btw. im a btech student in hyderabad doing an ops job side by side. started with graphic designing, then wordpress, then n8n, now building with ai picked pharmacy billing because most medical shops near me still use old desktop software or write bills by hand stack ended up as next js + react + typescript + tailwind, next api routes with better auth, postgres with prisma, and a separate python service for ocr using opencv and onnx. all in docker on a small ubuntu vps it does fefo stock (first expiry first out), batch and expiry tracking, returns, and bills can be shared on whatsapp. got cashfree gateway approval also. whole thing cost me around 3k in tools ai wrote most of the code honestly, i did the planning, testing and lots of fixing now its in testing at an actual pharmacy and scanning is the real problem. small print on strips, light glare, crushed boxes. on my table it works fine, at the counter totally different story anyone here done ocr or barcode for messy labels? did preprocessing in opencv help you more or switching the model?
document fetched 2026-10-31 00:25 · canonical https://reddit.com/r/developersIndia/comments/1xj4gte/spent_around_3k_rupees_on_tools_and_built_a_full