Skip to content

Use AI to generate llms.txt blurbs dynamically during file generation #908

@sirugh

Description

@sirugh

Summary

The llms.txt index file contains manually written blurbs that describe the contents of each generated bundle (e.g., in scripts/generate-llms-full.js). Since AI reads these blurbs in llms.txt to understand file contents, it's important they accurately reflect what each file contains.

Currently, these blurbs are hardcoded and risk becoming stale as content changes.

Proposed improvement

During the llms.txt generation process, use AI to dynamically generate accurate blurbs based on the actual content of each generated file — rather than relying on static, manually maintained descriptions.

This would keep blurbs in sync automatically and improve the quality of AI-readable metadata over time.

Context

Raised in PR #900 discussion — not a blocker for the current PR, but noted as a future improvement once AI can be hooked into the publication process.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions