@crag-pro/doc-parser

Multi-format document parser with OCR fallback. Extracts text from PDF, Word, Excel, PowerPoint, and plaintext files. Falls back to Tesseract OCR for scanned PDFs and an AI vision API for low-quality scans.

Originally built for CRAG (regulatory compliance intelligence) and extracted as a standalone library.

Install

npm install @crag-pro/doc-parser

Quick start

import { parse } from "@crag-pro/doc-parser";

const result = await parse("./contract.pdf");
console.log(result.text);
console.log(result.metadata);

Supported formats

Extension	Parser	Notes
.pdf	pdf-parse	OCR fallback (Tesseract) for scanned pages, AI vision API for low-confidence OCR
.docx	mammoth
.xlsx	exceljs
.pptx	XML extraction
.txt, .csv, .md, .msg	plaintext

CLI

npx @crag-pro/doc-parser ./input-dir ./output-dir

See doc-parser --help for full options.

OCR setup (optional)

Tesseract must be installed locally for OCR fallback:

# macOS
brew install tesseract

# Ubuntu
apt-get install tesseract-ocr

For AI vision fallback on low-quality scans, set ANTHROPIC_API_KEY in your environment.

License

MIT