Sarah MVP Development Roadmap

This will be my first project where I build it entirely myself with no AI coding assistance. After I have an MVP I built, and understand the mistakes I make, then I’ll reconsider using AI to maintain and expand upon my project.

Based on the roadmap and boot.dev assignment to spend 20-40 hours coding and implementing the software, the MVP might take the full timeframe. Once I have the core logic working, and everything behaves as indented (I’ll need to test several file versions), I’d like to focus on making the software a standalone application and polishing the UI.

UI has never been my focus before now, but this might be the right project to start. Might us AI to make the GUI after the fact. Not sure if I’d count that as cheating myself, as I’m not trying to learn/master front-end.

🧱 Phase 1: Core Architecture

1. Frontend (PyQt6)
   • File upload field
   • Multi-select checkboxes (e.g., grammar, tone, conciseness)
   • Text field: “Briefly describe what you’re aiming for”
   • Submit button
   • Status updates + file download link
2. Backend
   • Parse uploaded file
     • Use python-docx for .docx
     • Plain text reader for .txt, .md, etc.
   • Construct a prompt based on user input
   • Send prompt + file content to OpenAI API
   • Receive and write edited file to disk
   • Return download path to PyQt UI

🧰 Required Libraries

Frontend:

• PyQt6
• PyQt6.QtWidgets for the GUI elements

Backend:

• openai (API access)
• python-docx (for .docx reading/writing)
• mammoth (optional, for converting .doc to HTML/text)
• markdown (optional, if you want to clean up .md formatting)

🔧 Design Decisions

🚀 Initial Deliverables

• ✅ Basic PyQt6 app with file picker, options, and “submit”
• ✅ Backend script that:
  • Accepts a .txt or .docx
  • Accepts user options + brief
  • Sends content to OpenAI with a pre-written prompt
  • Returns a new file