OCR a document is part of the Documents category in the Workhog agent. When you pick it, your next chat message is wrapped so the agent uses this tool to run your command — no more guessing whether the agent will pick the right tool.
Prompt seed
Extract text from …your request herePDFs, scanned forms, and annotated documents are stored in your encrypted Workhog workspace. The agent can OCR a scan, extract structured text, fill form fields, or annotate pages — and the results are saved right back to your file system.
The agent sends your scanned document to AWS Textract for OCR, receives structured text blocks, and can write the extracted content back to your workspace. It handles multi-page scans, receipts, handwritten notes, and printed forms — anything Textract can read.
Once the agent OCRs a document or fills a form, the output is a workspace file you can revisit anytime. Future runs can reference that extracted text — the agent won't need to re-scan. Filled forms are saved as new PDFs, so your originals stay untouched and every version is retrievable.
Workflows where OCR a document is one step in a larger chain.
OCR a scanned form, pull the structured fields out, then auto-fill the matching PDF in your workspace.
Extract text from a scan, highlight or comment the key passages, and move the result into the right folder.