llms-keeper

To generate llms.txt files for AI agents, scan your software project to extract high-signal architecture and configuration data, then output both a navigation index and a full-context file following the llmstxt.org standard.

The llms.txt file serves as a concise navigation index for fast AI orientation, whereas the llms-full.txt file provides the complete, self-contained project context required for deep AI understanding.

You keep llms.txt documentation in sync by applying a workflow-driven update process that scans the repository, extracts high-signal information, and patches or regenerates the files to match the project state.

Yes, the llms.txt workflow applies directly to existing software engineering projects without requiring external dependencies, using a self-contained two-file structure to capture key patterns and data flows.

The best way to structure project context is enforcing a two-file structure that captures signal-first content like architecture and data flows, separating the navigation index from the complete project context.

You need to use llms.txt files when onboarding AI agents to a software project requires structured, high-signal context that accelerates their understanding of architecture, patterns, and configurations.