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