This software connects your local artificial intelligence models to your computer source code. Large language models often struggle to understand deep project structures or complex libraries. This tool solves that problem by processing repositories into structured data formats.
We scan thousands of public projects to create files that help your AI understand coding patterns. When your AI model needs to write code or solve problems, it uses these files to make better decisions. You process your data locally. This means your private code stays on your machine.
Your computer needs specific parts to run this software well. Please check your system against this list before you start.
- Operating System: Windows 10 or Windows 11.
- Processor: A modern multi-core processor from Intel or AMD.
- Memory: At least 16 gigabytes of RAM.
- Storage: 5 gigabytes of free space for the application and temporary data.
- Graphics: A dedicated graphics card with at least 8 gigabytes of video memory is recommended for faster performance.
You can download the application from our release page. Visit this link to see the available versions.
Download the latest version here
Look for the file that ends with a .exe extension. Click the file name to start the download. Save this file to a folder you can find easily, such as your Downloads folder.
- Locate the file you downloaded in your folder.
- Double-click the file to open the installer.
- Windows may show a security window. This is normal for new applications. If you see this, click "More info" and then "Run anyway."
- Follow the instructions on the screen to finish the setup.
- The installer places a shortcut on your desktop.
Using this tool involves three simple steps. You set your source folders, process the files, and connect your AI model.
Open the application from your desktop. The first window asks for the path to your code projects. Click the "Browse" button to select the folder where you keep your coding projects. The software scans this folder to index your files. It creates small data files that act as a map for your AI agent.
Once you select your folder, click "Build Index." The application transforms your code into JSON, Markdown, and Parquet files. This process may take a few minutes if you have many files. A progress bar shows you how much work is left. You do not need to repeat this step unless your code changes significantly.
The final step involves linking the software to your local AI model. Most users prefer tools that run models privately. In the settings panel, select your model provider. Enter the address where your model runs. If you use a tool that hosts models locally, it usually provides an address like http://localhost:11434. After you enter the address, click "Test Connection." A green checkmark confirms that your AI can now read the data files created in the previous step.
Most issues relate to file permissions or memory limits. If the software crashes, try the following steps.
- Check file access: Ensure the folder you picked is not a system folder. Windows restricts access to system folders for security reasons. Pick a folder in your Documents or a custom project folder.
- Manage memory: If the application closes during the indexing process, you likely have too many files selected. Select a smaller sub-folder to index instead of your entire hard drive.
- Update drivers: If the AI model fails to connect, make sure your graphics card drivers are current. Outdated drivers often cause issues with local AI processes.
- Check disk space: The tool creates several large files during indexation. Ensure you have enough disk space before you start the process.
Does this tool send my code to the internet? No. All processing happens on your local machine. Your code never leaves your computer.
Can I use this with any AI model? You can use this with most local models that support tool calling. The files we create are standard formats that work with several popular AI platforms.
How do I clear the cached data? Go to the "Settings" menu and find the "Clear Cache" button. This deletes the created JSON and Markdown files. Use this if you want to perform a clean scan of your project files.
How does this improve my AI results? Small local models often lack context. By providing them with a structured map of your code, you give them the information they need to write better logic and find bugs.
If you encounter a problem, open a new issue on our page. Include your Windows version and a description of the error you see. Our team reviews these reports to improve the software for everyone. Please maintain clear communication when reporting issues so we can reproduce the error on our machines.