home / skills / dkyazzentwatwa / chatgpt-skills / business-card-scanner

business-card-scanner skill

This skill extracts contact details from business cards using OCR, delivering name, company, email, phone, and address with structured exports.

npx playbooks add skill dkyazzentwatwa/chatgpt-skills --skill business-card-scanner

Review the files below or copy the command above to add this skill to your agents.

Files (3)

SKILL.md

736 B

---
name: business-card-scanner
description: Extract contact information from business card images using OCR - name, company, email, phone, address.
---

# Business Card Scanner

Extract contact information from business card images using OCR.

## Features

- **OCR Extraction**: Extract text from card images
- **Contact Parsing**: Name, company, email, phone, address
- **Pattern Recognition**: Smart regex for contact fields
- **Multi-Format**: JPG, PNG support
- **Batch Processing**: Multiple cards
- **Export**: vCard, JSON, CSV output

## CLI Usage

```bash
python business_card_scanner.py --input card.jpg --output contact.json
```

## Dependencies

- pytesseract>=0.3.10
- pillow>=10.0.0
- opencv-python>=4.8.0
- pandas>=2.0.0

Overview

This skill extracts contact information from business card images using OCR and targeted parsing rules. It recognizes names, company names, emails, phone numbers, and postal addresses, then exports structured contacts in vCard, JSON, or CSV formats. It supports JPG and PNG inputs and can batch-process multiple cards. The implementation is Python-based and optimized for practical, reliable extraction.

How this skill works

The skill runs OCR on input images to capture all visible text. It then applies pattern recognition and regex rules to identify and classify fields such as name, company, email, phone, and address. Multiple cards can be processed in a single run and results are normalized and exported to common contact formats. Optional pre-processing (deskewing, contrast) improves accuracy for low-quality images.

When to use it

Digitizing stacks of printed business cards after networking events
Populating CRM or address books from physical card collections
Automating contact entry for sales, recruiting, or partnership teams
Converting paper archives into searchable digital contact lists
Batch-processing vendor or attendee lists for conferences

Best practices

Use clear, well-lit photos or high-resolution scans for best OCR accuracy
Pre-process images to crop, deskew, and enhance contrast when cards are damaged or photographed at an angle
Review parsed results before bulk import to catch ambiguous fields or OCR mistakes
Standardize phone and address formats after extraction for consistent CRM imports
Run batch jobs during off-hours for large volumes to avoid rate limits or local CPU spikes

Example use cases

Take photos of 200 business cards after a trade show and export contacts as vCard for import into email clients
Automate daily extraction of new vendor cards and push JSON results into an internal contact API
Scan recruiting event cards and produce a CSV for the applicant tracking system
Convert legacy printed partner lists into a searchable address book with normalized phone and address fields

FAQ

What image formats are supported?

JPG and PNG are supported; results improve with higher resolution images.

How accurate is the extraction?

Accuracy depends on image quality and card design; clear text and simple layouts yield the best results. Pre-processing can boost performance.

What export formats are available?

Exports include vCard, JSON, and CSV for easy import into CRMs and address books.