Pdf Text ExtractorSAFE

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

# Install
clawhub install pdf-text-extractor

# Extract text from PDF
cd ~/.openclaw/skills/pdf-text-extractor
node index.js extractText '{"pdfPath":"./document.pdf","options":{"outputFormat":"text"}}'

const result = await extractText({
pdfPath: './invoice.pdf',
options: { outputFormat: 'text' }
});

console.log(result.text);

const result = await extractText({
pdfPath: './contract.pdf',
options: {
outputFormat: 'json',
includeMetadata: true
}
});

console.log(result.metadata);
console.log(`Words: ${result.wordCount}`);

const results = await extractBatch({
pdfFiles: [
'./doc1.pdf',
'./doc2.pdf',
'./doc3.pdf'
]
});

console.log(`Processed ${results.successCount}/${results.results.length} documents`);

const result = await extractText({
pdfPath: './scanned-doc.pdf',
options: {
ocr: true,
language: 'eng',
ocrQuality: 'high'
}
});

console.log(result.text);

const stats = await countWords({
text: result.text,
options: { countByPage: true }
});

console.log(`Total words: ${stats.wordCount}`);
console.log(`Pages: ${stats.pageCounts.length}`);
console.log(`Avg per page: ${stats.averageWordsPerPage}`);

const lang = await detectLanguage(text);

console.log(`Language: ${lang.languageName}`);
console.log(`Confidence: ${lang.confidence}%`);

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: pdf-text-extractor
description: Extract text from PDFs with OCR support. Perfect for digitizing documents, processing invoices, or analyzing content. Zero dependencies required.
metadata:
  {
    "openclaw":
      {
        "version": "1.0.0",
        "author": "Vernox",
        "license": "MIT",
        "tags": ["pdf", "ocr", "text", "extraction", "document", "digitization"],
        "category": "tools"
      }
  }
---

# PDF-Text-Extractor - Extract Text from PDFs

**Vernox Utility Skill - Perfect for document digitization.**

## Overview

PDF-Text-Extractor is a zero-dependency tool for extracting text content from PDF files. Supports both embedded text extraction (for text-based PDFs) and OCR (for scanned documents).

## Features

### ✅ Text Extraction
- Extract text from PDFs without external tools
- Support for both text-based and scanned PDFs
- Preserve document structure and formatting
- Fast extraction (milliseconds for text-based)

### ✅ OCR Support
- Use Tesseract.js for scanned documents
- Support multiple languages (English, Spanish, French, German)
- Configurable OCR quality/speed
- Fallback to text extraction when possible

### ✅ Batch Processing
- Process multiple PDFs at once
- Batch extraction for document workflows
- Progress tracking for large files
- Error handling and retry logic

### ✅ Output Options
- Plain text output
- JSON output with metadata
- Markdown conversion
- HTML output (preserving links)

### ✅ Utility Features
- Page-by-page extraction
- Character/word counting
- Language detection
- Metadata extraction (author, title, creation date)

## Installation

```bash
clawhub install pdf-text-extractor
```

## Quick Start

### Extract Text from PDF

```javascript
const result = await extractText({
  pdfPath: './document.pdf',
  options: {
    outputFormat: 'text',
    ocr: true,
    language: 'eng'
  }
});

console.log(result.text);
console.log(`Pages: ${result.pages}`);
console.log(`Words: ${result.wordCount}`);
```

### Batch Extract Multiple PDFs

```javascript
const results = await extractBatch({
  pdfFiles: [
    './document1.pdf',
    './document2.pdf',
    './document3.pdf'
  ],
  options: {
    outputFormat: 'json',
    ocr: true
  }
});

console.log(`Extracted ${results.length} PDFs`);
```

### Extract with OCR

```javascript
const result = await extractText({
  pdfPath: './scanned-document.pdf',
  options: {
    ocr: true,
    language: 'eng',
    ocrQuality: 'high'
  }
});

// OCR will be used (scanned document detected)
```

## Tool Functions

### `extractText`
Extract text content from a single PDF file.

**Parameters:**
- `pdfPath` (string, required): Path to PDF file
- `options` (object, optional): Extraction options
  - `outputFormat` (string): 'text' | 'json' | 'markdown' | 'html'
  - `ocr` (boolean): Enable OCR for scanned docs
  - `language` (string): OCR language code ('eng', 'spa', 'fra', 'deu')
  - `preserveFormatting` (boolean): Keep headings/structure
  - `minConfidence` (number): Minimum OCR confidence score (0-100)

**Returns:**
- `text` (string): Extracted text content
- `pages` (number): Number of pages processed
- `wordCount` (number): Total word count
- `charCount` (number): Total character count
- `language` (string): Detected language
- `metadata` (object): PDF metadata (title, author, creation date)
- `method` (string): 'text' or 'ocr' (extraction method)

### `extractBatch`
Extract text from multiple PDF files at once.

**Parameters:**
- `pdfFiles` (array, required): Array of PDF file paths
- `options` (object, optional): Same as extractText

**Returns:**
- `results` (array): Array of extraction results
- `totalPages` (number): Total pages across all PDFs
- `successCount` (number): Successfully extracted
- `failureCount` (number): Failed extractions
- `errors` (array): Error details for failures

### `countWords`
Count words in extracted text.

**Parameters:**
- `text` (string, required): Text to count
- `options` (object, optional):
  - `minWordLength` (number): Minimum characters per word (default: 3)
  - `excludeNumbers` (boolean): Don't count numbers as words
  - `countByPage` (boolean): Return word count per page

**Returns:**
- `wordCount` (number): Total word count
- `charCount` (number): Total character count
- `pageCounts` (array): Word count per page
- `averageWordsPerPage` (number): Average words per page

### `detectLanguage`
Detect the language of extracted text.

**Parameters:**
- `text` (string, required): Text to analyze
- `minConfidence` (number): Minimum confidence for detection

**Returns:**
- `language` (string): Detected language code
- `languageName` (string): Full language name
- `confidence` (number): Confidence score (0-100)

## Use Cases

### Document Digitization
- Convert paper documents to digital text
- Process invoices and receipts
- Digitize contracts and agreements
- Archive physical documents

### Content Analysis
- Extract text for analysis tools
- Prepare content for LLM processing
- Clean up scanned documents
- Parse PDF-based reports

### Data Extraction
- Extract data from PDF reports
- Parse tables from PDFs
- Pull structured data
- Automate document workflows

### Text Processing
- Prepare content for translation
- Clean up OCR output
- Extract specific sections
- Search within PDF content

## Performance

### Text-Based PDFs
- **Speed:** ~100ms for 10-page PDF
- **Accuracy:** 100% (exact text)
- **Memory:** ~10MB for typical document

### OCR Processing
- **Speed:** ~1-3s per page (high quality)
- **Accuracy:** 85-95% (depends on scan quality)
- **Memory:** ~50-100MB peak during OCR

## Technical Details

### PDF Parsing
- Uses native PDF.js library
- Extracts text layer directly (no OCR needed)
- Preserves document structure
- Handles password-protected PDFs

### OCR Engine
- Tesseract.js under the hood
- Supports 100+ languages
- Adjustable quality/speed tradeoff
- Confidence scoring for accuracy

### Dependencies
- **ZERO external dependencies**
- Uses Node.js built-in mo

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

not all pinned

Secrets in source

none-found

Findings (2)

<key>k5Ey9KFlkqpj+SDkUw+5ED9lTA3En/qUi0zdrydUCH3kMWTE3Eh65NXnFCaxlY2omY2JHnlEoK7Li7oOEvM7eG5VPdcO/sFlMfoCRdnLYdepJ+uLzYwOWR8W4yQVve/clxVFTVRL4DFleKInGdpAxIbHZT2yi4ADAMENls1N1XSLojRuqXePXDeAT/4Mv4TTx0s

pdfjs-dist

Gates applied: no_behavioural_pass.