mcpField
O
ocr-document-processor

ocr-document-processor

github.com/dkyazzentwatwaDocuments1244.8k installsUpdated 29w ago
Repository

About ocr-document-processor

ocr-document-processor is a skill for recovering and structuring text from images and scanned documents. It covers general OCR as well as receipt and business-card parsing, with guidance for handling low-quality sources.

Key features

  1. 01OCR for images and scanned PDFs
  2. 02Searchable PDF export
  3. 03Extraction to text, Markdown, JSON, or HTML
  4. 04Table extraction
  5. 05Receipt and business-card parsing

Use cases

  1. 01Make scanned PDFs searchable
  2. 02Extract tables from scanned pages
  3. 03Parse receipts or business cards

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/dkyazzentwatwa/chatgpt-skills --skill ocr-document-processor

After installing, type/ocr-document-processorin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

.DS_Store
agents/
openai.yaml
scripts/
business_card_scanner.py
ocr_processor.py
receipt_scanner.py
requirements.txt

Details

From the project README

OCR Document Processor

Handle OCR-heavy inputs where text must be recovered from images or scanned pages.

Use This For

  • OCR on images and scanned PDFs
  • Searchable PDF export
  • Structured extraction to text, markdown, JSON, or HTML
  • Table extraction from scanned material
  • Receipt parsing and business card parsing

Workflow

  1. Decide whether plain OCR, structured extraction, or document-specific parsing is needed.
  2. Preprocess noisy inputs before extraction when skew, blur, or shadows are present.
  3. Use scripts/ocr_processor.py for core OCR tasks.
  4. Use the focused helpers when the input is specialized:
    • scripts/business_card_scanner.py
    • scripts/receipt_scanner.py
  5. Return confidence caveats when the source is low quality, rotated, handwritten, or multilingual.

Guardrails

  • Prefer explicit language selection when accuracy matters.
  • Do not claim fields are exact when OCR confidence is weak.
  • Route non-scanned digital PDFs to document-converter-suite instead of OCR by default.

Comments

Markdown supported