The Apache Tika Content Extraction Hub provides universal document parsing through the Apache Tika Server REST API. It extracts text, metadata, and structure from over 1400 file formats including PDF, Microsoft Office (DOCX, XLSX, PPTX), email archives (MBOX, PST), and multimedia files.

The skill communicates with Tika Server endpoints including /tika for text extraction, /meta for metadata parsing, /detect/stream for MIME type detection, and /rmeta for recursive metadata extraction from container formats. Batch processing supports concurrent extraction with configurable thread pools.

Advanced PDF extraction uses Tika PDFParser with OCR fallback via integrated Tesseract for scanned pages. Embedded document extraction recursively processes attachments within ZIP archives, email messages, and OLE containers, maintaining parent-child document relationships.

Metadata extraction captures Dublin Core, XMP, EXIF, and format-specific properties with normalization to a common schema. The agent supports custom Tika configurations for parser-specific settings like OCR language, PDF strategy (text vs OCR vs hybrid), and maximum extraction depth. Output formats include plain text, XHTML, and JSON with full metadata. Language detection via Tika LanguageIdentifier classifies extracted content across 70+ languages.