I need a lean proof-of-concept that lets a smallholder farmer open a web page, choose a crop image from their phone or computer, and receive an instant, AI-generated diagnosis plus concise treatment advice. Core requirements • Single-page web app (mobile-friendly). • Simple file-upload widget for the photo; no live camera capture at this stage. • Back-end endpoint loads my pre-trained model (Python, TensorFlow/Keras or PyTorch—whichever you prefer) and returns: – disease or pest identified, – severity score, and – short, actionable recommendations. • Optional field for the farmer to attach a small CSV or Excel file with local weather or soil readings; your script should parse the sheet, append the values to the inference request, and adjust the advice accordingly. • Clean JSON logging so the data can later feed a monitoring dashboard. Keep the UI vanilla (HTML, CSS, a lightweight JS framework is fine) and concentrate your time on wiring the upload, model inference, and response formatting. A brief README with setup steps and sample input files will complete the delivery.