sakshamsinghania/ScholarScan — reverse-engineered prompt
Reverse engineered prompt
Build me a simple web app called ScholarScan for teachers to grade handwritten or scanned student answers.
I want two main flows. First, a quick manual grading mode where a teacher uploads one answer image, pastes the correct answer, and gets a score plus readable feedback right away. Second, a document grading mode where the teacher uploads a full answer sheet as an image or PDF, can also attach the question paper, and the app extracts the content, figures out the questions and answers, grades them in the background, and shows live progress until the final result is ready.
Please include login and registration, basic teacher friendly access control, saved assessment history, and a clean dashboard for uploads, progress, results, and failed jobs. It should handle noisy handwriting and scans as well as possible, and give useful feedback about what the student covered and what they missed. If document grading cannot confidently ground answers in the actual question text, it should refuse to score instead of guessing.
Use the existing repo structure and look up current docs online if you need to.
Want more depth? Deep Reverse