AI Code Prep GUI - Vibe Code Faster

This SAVES YOU TIME! No more copy paste copy paste!

We have lots of tools to code with AI today.. like Cursor, Cline, Windsurf, v0, but sometimes they do not give the AI the right context. Often way too little or too much causing the LLM to get confused. 

Then you are like "ah.. ugh" and have to copy lots of code to a AI chat (Gemini in AI studio, ChatGPT, Openrouter, Claude 3.5 or 3.7 Sonnet, etc) and you have to copy and paste each file over and over, it sucks. That's why I created this little app.

Windows GUI application (also have Mac and Electron/Linux) designed to streamline the process of sharing your project's code with ChatGPT, Claude, other AI chatbots. It allows you to quickly gather code files into a single text file and copy the content to your clipboard, making it easy to paste into AI chatbots for coding assistance. Uses xml (or markdown) tags that AI understands.

Purpose

The main goal of aicodeprep is to save time, and save on tokens. Instead of copying and pasting over and over for each code file, it auto-detects most likely needed files, and lets you check or uncheck any other files to add/subtract. With the GUI, you can effortlessly right click in a folder, select/unselect files, then process to clipboard/txt file with automatic tags and filenames in the text.

Features

Right-Click Context Menu: After installation, you can right-click in any Windows folder, click AI Code Prep GUI, to open the GUI and auto-detect code files.

It auto-skips the ones that you probably don't need to copy, like node files or .min.css, git files, etc.

Easy Processing: Click "Process Selected" to compile code into fullcode.txt and copy it to the clipboard.

This version is lean but if you want a different version that is more polished looking, look on my profile for the Electron versions. Also have a Mac version.

-- Buying this will get you the newest Windows version, the older Windows version (just in case the updated one glitches or something), and a Linux version
-----

It may not look sophisticated, but a lot of other tools will burn though way too many tokens from including too much information in each API call. Claude 3.5 Sonnet will cut you off if you use too many too fast. This allows fine control and it saves you money. It's designed for speed, something I personally use every day.
-----
When you use the app and click some files, and click Process Selected, the output to clipboard and fullcode.txt looks like this:

Example output:
----------------------------

App.css:
<code>
..code here..
</code>

App.jsx:
<code>
..code here..
</code>

firebase.js:
<code>
..code here..
</code>

fileName.cpp:
</code> 
code from file here...
</code>

Text from prompt box here...

----------------------------

Then you can quickly paste it right into Gemini on AI studio, ChatGPT, Deepseek, etc.

Made a Mac version and also a Web/Electron version.
<ai> <code> <aicodeprep> <claude> <ChatGPT> <llm> <deepseek> <openai> <vibe coding> 

### Updated April 30 2025...

-  Collapsable/Expandable folders, automatic expand if code detected

-  Token counter (live, as you click files)

-  Better smart file automatic pre-checking on load

-  Saves an .aicodeprep file with the files you last had checked in that folder, plus your window size / adjustment preferences. This helps when you need to keep grabbing the code / context often for the same project. I hated having to click the same boxes over again so it's a great feature.

- Optional prompt box to type in something such as a question to ask AI - it just appends this to the bottom of the code context section - which will be on the clipboard and fullcode.txt

- Sorry Mac version hasn't been compiled yet, will soon. This zip contains the newest Windows version, the older Windows version (in case something goes wrong, at least we know the older one is very stable), and a Linux version all in one zip file. Standard Windows installer wizards.

- Also included a html guide/writeup of how to vibe code or use AI to help code, using the best models but still being cheap, even free. It took a while to figure all that stuff out on there so it should save anyone some time. What models to use when, what do you do if you want to save money, how to debug free and which web chat's you can use free/unlimited. As of today, your options are wild 😂


----