auto-prime

What problem does it solve?

Automates turning unprocessed client documents into structured RAG priming files with evidence tagging.

Core Features & Use Cases

• Read unprocessed client documents in various formats (PDF, HTML, Word) and extract structured knowledge.
• Classify documents by type (textbook, curriculum, inserts, deliverables, taxonomy, drafts) and generate per-document priming files with frontmatter including source, type, domain, processed_date, and confidence, plus TL;DR, Key Entities, Constraints, and Evidence.
• Save priming files to references/priming-rag/ and update calibration-digest.md, then report processed counts and new calibration level for validation.

Quick Start

Instruct the AI to read unprocessed client documents and generate corresponding RAG priming files.