Apache Tika
CertifiedTasks that extract text and metadata from files using Apache Tika.
Provide input files to detect content type and extract text or metadata across many formats (PDF, Office, images, etc.), returning parsed content for further processing in workflows.
Apache Tika
Tasks that extract text and metadata from files using Apache Tika.
Provide input files to detect content type and extract text or metadata across many formats (PDF, Office, images, etc.), returning parsed content for further processing in workflows.
How to use the Tika plugin
Extract text content and metadata from documents (PDF, Word, Excel, images, and more) using Apache Tika from Kestra flows.
Tasks
Parse parses a document and extracts its text and metadata — set from (required, a kestra:// URI). Control output format with contentType (XHTML by default; also TEXT or XHTML_NO_HEADER). Set extractEmbedded: true to also extract embedded documents (default false). Set charactersLimit to cap extracted text length. Set store: false to return results inline rather than writing to internal storage (default true).
Configure OCR via the ocrOptions object: set strategy (default NO_OCR; set to OCR_ONLY or OCR_AND_TEXT_EXTRACTION to enable Tesseract), language (Tesseract language code), and enableImagePreprocessing.
When store: true (default), the output includes uri (ION file in internal storage). When store: false, the output includes result with content (extracted text), metadata (map of document properties), and embedded (map of embedded document name → URI).
