The Scratch: Note Compiler

The Scratch: Note Compiler

Turn messy notes into clean writing — locally, offline, and on your terms.

The Scratch is a local-first note compiler built for students who actually think.

You paste in rough notes — class notes, lecture dumps, study guides, half-baked ideas — and The Scratch turns them into usable outputs without pretending to be a chat buddy or a cloud service.

No accounts.

No subscriptions.

No internet required (after setup).

Just input → output.

What it does

The Scratch takes your raw notes and compiles them into:

• Summaries – tight, readable overviews
• Essays – expanded, structured writing
• Outlines – clean, logical structure
• Key Points – extracted facts and ideas (not filler)

You control how compressed or expanded the output is.

No guessing. No “AI vibes.” Just results.

Built for students

This tool is especially good for:

• High school & college students
• STEM, humanities, and research-heavy classes
• Studying from messy lecture notes
• Turning bullet points into papers
• Cleaning up thoughts before writing assignments

It’s fast, boring, and reliable — in a good way.

Anti-AI Mode (yes, really)

The Scratch includes an Anti-AI / Plain mode that:

• Avoids obvious AI phrasing
• Suppresses filler language
• Keeps your original keywords and voice

Your output sounds like you, not a chatbot.

How it works

• Runs 100% locally on your computer
• Uses a local LLaMA model via Ollama
• Outputs plain text files (.txt / .md)
• No memory, no tracking, no cloud

You own everything.

Requirements (important, read this)

• macOS, Windows, or Linux
• Python installed
• Ollama installed with a LLaMA model
• (Setup instructions included — no coding required)

This is not a web app and not a mobile app.

It’s a local tool for people who want control.

Price

$1.99 one-time purchase.

No upgrades, no subscriptions, no nonsense.

If you’re a student and you want a tool that helps you think and write better — this is for you.

Required (one-time setup): 

• Python 3.11 or newer (used to run the application locally). Download: https://www.python.org/downloads/ — make sure Python is added to your system PATH during install.
• Ollama + LLaMA model (used to run the AI locally; nothing goes to the cloud). Download Ollama: https://ollama.com/ then after installing, run ollama pull llama3.2.