Speed of Cursor?

Feedback
1

First of all, I am building my first project with Cursor. I have subscribed to Cursor Pro. I am working on a PDF renaming desktop app. I want to rename PDFs based on the information that is contained within them. We have uniform formats that will have the info within the Docs. I had Gemini help me with the Prompt below. I copy and pasted into the Chat window about 16 hours ago and cursor is still spinning to write the code. Did I ask too much in one shot or are my expectations too high?

“Persona and Goal

You are a senior Python developer specializing in building cross-platform desktop applications. Your task is to build a self-contained, executable desktop application for a customer. You will implement the user interface, file handling, PDF text extraction, and file renaming logic exactly as described. You will not make any assumptions or deviate from the provided requirements.

Project Stack & Libraries

  • Primary Language: Python

  • Desktop Framework: [Suggestion: Use PyQt or Tkinter for a simpler UI. PyQt is recommended for a cleaner, more modern look.]

  • PDF Library: PyMuPDF or pdfplumber to extract text from the PDF. You will analyze the document to determine which library is most suitable.

  • File Handling: Standard Python libraries (os, shutil, re) for file operations and regular expressions.

  • Packaging: PyInstaller to create a single-file executable.

Core Functionality

The application will have the following user flow and features:

  1. Loading Screen:

    • On startup, clear the screen.

    • Display “Warr Machine-PDF Convertor. Created by RPJ Industries, Inc.”

    • Display the company logo below the text. (You will need to provide the logo file or a placeholder URL for the AI to reference).

    • Show a spinning wheel or progress indicator.

    • This screen will transition to the main menu when the application is ready.

  2. Main Menu:

    • Display “Warr Machine-PDF Convertor. Created by RPJ Industries, Inc.” and the logo at the top of the screen.

    • Show three buttons centered on the page: “Preview,” “Claim,” and “Credit Memo.”

    • Clicking any of these buttons will clear the screen and proceed to the next step.

  3. File Input Screen:

    • Display “Warr Machine-PDF Convertor. Created by RPJ Industries, Inc.” and the logo at the top of the screen.

    • Present a box where users can drag and drop a single PDF file.

    • Provide a button labeled “Select File” to open a file browser for uploading a document.

    • A “Rename” button will be present but will only become active once a file is loaded.

    • The renaming process will begin when the user clicks “Rename” or hits the “Enter” key after a file is loaded.

  4. Renaming Logic:

    • Based on the user’s initial button selection, the program will perform the following searches within the uploaded PDF’s text:

    • If ‘Preview’ was selected:

      • Search for “Key Part #:” and “Work Order #:”

      • Rename format: key part#_work order#_Preview.pdf

    • If ‘Claim’ was selected:

      • Search for “Failed Part:” and “Work Order 1:” and “Claim Number:”

      • Rename format: Failed Part_Work Order1_Claim Number_CL.pdf

    • If ‘Credit Memo’ was selected:

      • Search for “Failed Part:” and “Dealer Work Order 1:” and “John Deere Claim No:”

      • Rename format: Failed Part_Dealer Work Order1_John Deere Claim No_CM.pdf

    • You will use regular expressions to accurately extract the values following these labels, ensuring they are correctly captured for the new filename.

    • The final renamed file will be saved to the user’s desktop Downloads directory.

  5. Completion Screen:

    • Clear the screen and display the same header and logo.

    • Display the message “Renaming is complete.”

    • Show the newly named PDF file on the screen.

    • The user should be able to drag this file directly from the application’s UI to a folder on their desktop.

    • Two buttons will be displayed at the bottom of the screen:

      • “Next Document” will return the user to the main menu (step 2).

      • “Quit” will exit the application.

Implementation Plan

Your first step is to set up the project structure and create the necessary boilerplate for a standalone desktop application using Python. Please create the following:

  • A main Python file (app.py).

  • A basic requirements.txt listing the necessary libraries (PyQt, PyMuPDF, PyInstaller).

  • A simple, functional UI for the loading screen (step 1) and the main menu (step 2) to demonstrate the application’s basic flow.

  • Include placeholders for the logo and the UI elements for the subsequent steps.

  • Use clear, well-commented code that adheres to standard Python best practices.

2

Which model are you on?

3

1.4.2

4

This is what model your requests are going to. Which model are you using?

image
image332×130 3.48 KB

5

It says “balanced”. It doesn’t really indicate any specific model. I think I am going to “scrap” the session and start over. I have inquired at Github and a couple other forums and everything is leading toward an issue that I may have created. Is there anything obvious that I am overlooking before I scrap it?

6

You may want to start with something smaller. Just to get your bearings with programing and using these tools. Instead of having Cursor just create an entire project in one request. There is a learning curve and most likely it wont work on the first go so if you don’t know how to give the model direction based on the results which most likely will not work as expected, it’s kind of pointless. AI is super helpful, just don’t skip learning some essentials as well. So start off with a small python script that just renames a PDF to a static word. Then add on from there. Good luck.

1 Like
7

Listen, my friend, you really sound like a beginner.
Most likely, you got stuck in a loop.
No normal model runs for that long continuously.
And this model wasn’t designed for that.

8

It would be worth your while to learn a bit about AI models — how to use them in Cursor and in other IDEs.
Pick a different model, run one small feature at a time, check that it works, and then move on.

And really, it might be better to choose a language like Next.js.
From the way your knowledge sounds, that would probably suit you much more.
Next.js and React can work even for someone who isn’t a programmer at all — as long as the right model is chosen.

9

Update. I cleared out of the session with the spinning circle and started over. Worked on prompting for about 4 hours and now have a working desktop app. Thanks for your input

10

For a request like this, Claude 4.1 Opus Thinking would be fine if you send everything in one request. If you want to save money, use Claude Sonnet 4 Thinking. The models in Auto are only suitable if you send each request separately.