Data Science & Analytics

OCR Document Processing Skill

Amazon Textract document processing without an MCP server. Extracts reading-order linearized text, raw OCR with low-confidence flagging, tables (Markdown/CSV), form key-value pairs, QUERIES answers, receipts/invoices (AnalyzeExpense), identity documents (AnalyzeID), per-block geometry and confidence, and the auto-accept/human-review confidence-routing pattern. Use for any OCR or document-understanding task on a PDF or image — "extract text", "reading order", "tables to markdown/CSV", "form fields", "ask this document a question", "parse this receipt", "driver's license fields", "process a 40-page PDF", "batch these documents", or "auto-accept high-confidence fields and flag the rest". Drives the AWS CLI directly (sync plus async polling) and runs a bundled parser (scripts/parse.py). Works in any runtime with AWS credentials (Textract + S3) and Python 3.10+. Also known as intelligent document processing (IDP): handles scanned PDFs, handwriting, and W2, KYC, passport, and claim documents.

Publisher / source
aws/agent-toolkit-for-aws (found on skills.sh)
Upstream record
aws/agent-toolkit-for-aws
Version
Version 1
Published
Publication review
Approved for publication
Content fingerprint
ed7b26c3ed2710ac2319ffe4615ca4bd1191a7628fd3aab9b28faa37171b0950