Astro Integration
Astro LLM
Extract deterministic, LLM-ready content from Astro builds at build time — no runtime JavaScript, servers, or magic.
Usage snapshot
Why this plugin exists
Large Language Models require clean, predictable, auditable source material. Runtime scraping and crawling introduce non-determinism and risk.
Astro LLM extracts readable content directly from your built HTML output in DOM order after astro build completes.
All behaviour is controlled by a single configuration file. Given the same input, the output is always identical.
What it delivers
Design principles
- Build-time only — runs after astro build
- Deterministic output — same input, same output
- Config-first behaviour via llm.config.json
- Safety-by-default content stripping
- LLM-friendly, auditable structure
What this plugin does
- Scans generated HTML files in /dist
- Extracts readable content in DOM order
- Applies include and exclude rules explicitly
- Strips sensitive data such as emails and phone numbers
- Outputs a single TXT or JSON file
Installation
npm install astro-llm
On first run, llm.config.json is created in the project root with explicit defaults and is never overwritten.
Project links
Source code, releases, documentation, and contribution guidelines.
Want the deep dive?
Read the FAQs for implementation details, design rationale, and integration guidance.
View plugin FAQs