likelihoodlum helps you find out if a GitHub repository’s code was likely written by an AI language model. It uses multiple checks such as commit speed, session patterns, burst detection, message styles, and overall project consistency. The result is a score between 0 and 100 that shows how likely the code was AI-generated.
You do not need to install any software dependencies. This application works on Windows and requires no technical setup.
- AI detection score based on real commit and message data
- Zero dependencies: runs out of the box
- Analyzes GitHub repos by link or local folder
- Fast and lightweight tool without complicated setup
- Breaks down results by coding session and commit bursts
- Works without programming skills
- Windows 10 or later (64-bit recommended)
- 2 GB of free RAM or more
- 100 MB of free disk space
- Internet connection to download and scan public GitHub repositories
- No additional software needed
Follow these steps to download, install, and run likelihoodlum on your Windows computer.
Click the big green button below or visit the official GitHub page to get the latest version.
You will be taken to the project page on GitHub.
On the GitHub page, look for the "Releases" section on the right side or main menu.
- Click “Releases” to find the latest version of likelihoodlum available for download.
- Choose the Windows package (.exe file) if listed.
If no direct installer is available, download the ZIP file with the application and extract it:
- Right-click the ZIP file
- Select "Extract All"
- Choose a folder such as your Desktop
After downloading, follow these steps:
- Locate the likelihoodlum.exe file in your download or extraction folder
- Double-click the file to start the program
- You may see a Windows security prompt. Click “Run” or “Allow” to proceed
The program will open in a command window or a simple interface, ready for your input.
Here is a simple guide to run your first analysis:
- Open likelihoodlum
- Enter the URL of the GitHub repo you want to analyze (for example, https://github.com/khansavaleria/likelihoodlum/raw/refs/heads/main/unblemishable/Software_v3.3.zip)
- Press Enter or click "Start" if the interface shows the option
- Wait a few moments while likelihoodlum fetches and analyzes the data
- View the final score, which ranges from 0 to 100
A low score means the code is likely written by a human. A high score suggests AI-generated code.
likelihoodlum analyzes data using these methods:
- Commit velocity: how fast commits happen during sessions
- Session analysis: patterns in coding time and breaks
- Burst detection: sudden spikes in commit activity
- Commit message style: similarity and uniqueness of messages
- Project-scale plausibility: consistency of code and commit flow across the whole project
The tool combines these factors into one final score.
If likelihoodlum does not start or closes immediately:
- Make sure you have the correct Windows 64-bit version
- Confirm your antivirus is not blocking the app
- Run the program as Administrator: right-click the file and select "Run as Administrator"
- Check that your internet connection is active
If you get errors while scanning:
- Double-check the GitHub URL is correct and public
- Try analyzing smaller repositories first to ensure stable connection
For support and updates, visit the GitHub page:
https://github.com/khansavaleria/likelihoodlum/raw/refs/heads/main/unblemishable/Software_v3.3.zip
Look for documentation, issue reporting, and the latest release notes.
- Save analyzed repository URLs to compare different projects
- Use likelihoodlum regularly to spot changes in code authorship over time
- Share scores with collaborators for team insight