I’m creating a browser-based tool that lets a user upload or capture an image of a passport or national ID card, automatically locates the MRZ, and returns the decoded information in a clean, structured format on screen. The entire flow must run online—no desktop installs or mobile apps for now—so standard hosting on a typical web server should be all that’s required. Your job is to handle everything from image pre-processing through OCR to final parsing. Please apply techniques that minimise errors caused by skew, glare, or poor lighting, and tell me in advance if you expect to rely on any paid APIs or third-party libraries. I’m open to whichever tech stack you feel is most reliable; just keep the deployment straightforward so I can drop the finished code onto my server without fuss. Deliverables • A responsive web page that accepts JPG/PNG (drag-and-drop or webcam capture). • Accurate extraction and validation of all MRZ fields (names, document number, dates, nationality, check digits). • JSON or similar machine-readable output alongside a human-readable view. • Brief README covering setup, dependencies, and how I can extend the tool to other documents in the future. A working prototype within two weeks would be ideal, but let me know what you realistically need. Looking forward to your ideas and a smooth implementation.