Document Parser

Gives Claude five document parsing tools: parse_pdf for extracting text and tables from PDFs with layout preservation, parse_image_text for OCR with confidence scoring across 100+ languages, html_to_markdown for clean conversions, extract_tables for pulling structured data from any format, and summarize_document with configurable detail levels. Built by Agenson Horrowitz with a freemium model starting at 500 operations per month. Useful when you're building agents that need to ingest reports, invoices, screenshots, or web pages and want structured output without managing separate parsing libraries. All responses come back as JSON with metadata like processing time and confidence scores.