GitHub Repository Synchronization

The GitHub Repository Synchronization feature is a powerful extension of the Antimanual Knowledge Base, designed specifically for technical teams and developers. By connecting your GitHub account directly to the plugin, you can transform your existing technical documentation, README files, and markdown notes into a high-performance AI knowledge source. This integration ensures that your AI chatbot remains up-to-date with your latest code-base documentation without the need for manual copy-pasting or file uploads.

• Overview of GitHub Synchronization
• Setup and Synchronization Workflow
• Technical Ingestion Details
• Best Practices for Repository Structure
• Frequently Asked Questions

Overview of GitHub Synchronization

For many software-driven organizations, the primary source of truth resides within a repository. Antimanual Pro allows you to tap into these repositories to train your AI on complex technical documentation. This feature is particularly valuable for creating developer-focused assistants that can answer questions about API endpoints, installation procedures, and architectural overviews directly from your project’s source files.

To get started with this feature, ensure you have already configured your to allow the AI to process embeddings for your new data.

Setup and Synchronization Workflow

The process of synchronizing a GitHub repository is streamlined into three primary steps within the Antimanual dashboard:

1. Account Connection: Authenticate your GitHub account within the plugin settings. This grants Antimanual the necessary permissions to read the repository list and file contents.
2. Repository Selection: Browse your available repositories and select the specific projects you wish to index. You can choose from both public and private repositories depending on your access level.
3. Indexing and Ingestion: Once selected, the AI will begin crawling the repository, specifically looking for README files and other markdown documentation.

Technical Ingestion Details

Antimanual focuses on markdown-based documentation to provide the highest quality context for the AI. When the synchronization begins, the system performs the following actions:

• File Filtering: The crawler prioritizes .md and .mdx files, ensuring that noise from raw code (like C++ or JavaScript) is minimized while the documentation is emphasized.
• Text Extraction: The content is stripped of unnecessary metadata and parsed for its structural hierarchy (headings, code blocks, and lists).
• Vectorization: The extracted text is sent to your selected AI provider to be converted into embeddings, which are then stored in your local Knowledge Base.

This process is managed within the broader Managing User Roles and Access in Docy: Best Practices and Plugin Recommendations system, where you can see the status of your indexed repository items.

Best Practices for Repository Structure

To maximize the accuracy of your AI responses when using GitHub as a source, consider the following organizational tips:

• Detailed READMEs: Ensure your root README.md contains a comprehensive overview of the project.
• Structured Docs Folder: Use a dedicated /docs folder to house specific guides, as this helps the crawler identify high-value information.
• Clear Headings: Use semantic H1, H2, and H3 tags within your markdown files. The AI uses these headers to understand the context of the information beneath them.

Frequently Asked Questions

Does Antimanual support private repositories?

Yes, as long as the authenticated GitHub account has read access to the private repository, Antimanual can index the markdown files contained within it.

How often does the repository sync?

Currently, the synchronization is a manual trigger. When you update your documentation on GitHub, you should return to the Antimanual Knowledge Base and refresh the specific repository source to pull the latest changes.

Can I index code files like .js or .py?

Antimanual is optimized for documentation rather than raw code analysis. While it primarily indexes markdown files, it will capture code snippets located within those markdown files to provide context for technical queries.