Distill Book

Build a reliable mental model of a book in minutes. Optimize for understanding per minute, not coverage per page.

Inputs

Accept:

• EPUB files
• Text-based or scanned PDF files
• Plain text, Markdown, HTML, or DOCX when the environment can extract them
• A book title only when the user explicitly asks for a synthesis from publicly available sources; disclose that this is not a source-grounded extraction of the full book

Match the user's language unless they request another language.

Workflow

1. Establish the reading goal