The API is running. Use the /extract-questions endpoint to extract questions from a PDF & /result endpoint to evaluate answers.