​
Overview

Pulse API supports a wide range of document formats, enabling extraction from virtually any business document. All file types are processed with the same high accuracy and can be used with structured schema extraction.

​
Supported Formats

PDF Documents

Extension: .pdf
  • Text-based PDFs (searchable)
  • Image-based PDFs (scanned)
  • Mixed content PDFs
  • Password-protected PDFs (provide password)
  • Multi-page documents

Images

Extensions: .jpg, .jpeg, .png
  • High-resolution scans
  • Photographs of documents
  • Screenshots
  • Multi-page TIFFs
  • Handwritten content (with limitations)

Office Documents

Extensions: .docx, .pptx, .xlsx
  • Microsoft Word documents
  • PowerPoint presentations
  • Excel spreadsheets
  • Embedded images and charts
  • Complex formatting preserved

Web Documents

Extensions: .html, .htm
  • Static HTML pages
  • Saved web pages
  • HTML emails
  • Inline styles preserved
  • Embedded images extracted

​
Processing Large Files

For very large files, we recommend:
  • Using async processing (/extract_async)
  • Processing specific page ranges
  • Contacting support for optimization strategies

​
Format-Specific Features

​
PDF Processing

PDFs receive the most comprehensive processing: 
# Extract with PDF-specific options
result = client.extract(
    file_path="document.pdf",
    pages="1-10",  # Process specific pages
    extract_figure=True,  # Extract embedded images
    return_html=False  # Get markdown output
)
Advanced PDF Features:
  • Preserve document structure
  • Extract form fields
  • Handle multi-column layouts
  • Process rotated pages
  • Extract embedded images

​
Image Processing

Images are processed using advanced OCR: 
# High-quality image extraction
result = client.extract(
    file_path="scanned_document.jpg"
)
Image Best Practices:
  • Resolution: Minimum 150 DPI, recommended 300 DPI
  • Format: PNG for text, JPG for photos
  • Size: Keep under 10MB per image
  • Quality: Avoid blurry or skewed images

​
Office Document Processing

Office files maintain their structure: 
# Extract structured data from Excel
schema = {
    "headers": ["string"],
    "data": [["string"]]
}

result = client.extract(
    file_path="spreadsheet.xlsx",
    schema=schema
)
Office Document Features:
  • Preserve table structures
  • Extract embedded objects
  • Handle multiple sheets/slides
  • Maintain formatting context

​
File Upload Methods

​
Direct Upload

Upload files directly in your API request: 
# Direct file upload (recommended for files < 10MB)
with open("document.pdf", "rb") as f:
    response = requests.post(
        "https://dev.api.runpulse.com/extract",
        headers={"x-api-key": API_KEY},
        files={"file": f}
    )

​
Pre-upload to S3

For larger files or reusable uploads: 
# Upload to S3 first
upload_response = requests.post(
    "https://dev.api.runpulse.com/convert",
    headers={"x-api-key": API_KEY},
    files={"file": open("large_file.pdf", "rb")}
)

file_url = upload_response.json()["s3_object_url"]

# Then extract using the URL
extract_response = requests.post(
    "https://dev.api.runpulse.com/extract",
    headers={"x-api-key": API_KEY},
    json={"file-url": file_url}
)

​
URL-Based Processing

Process files from public URLs: 
# Extract from URL
response = requests.post(
    "https://dev.api.runpulse.com/extract",
    headers={"x-api-key": API_KEY},
    json={"file-url": "https://example.com/document.pdf"}
)

​
Unsupported Formats

The following formats are not currently supported:
  • Video files (.mp4, .avi, .mov)
  • Audio files (.mp3, .wav)
  • CAD files (.dwg, .dxf)
  • Legacy Office formats (.doc, .xls, .ppt)
  • Compressed archives (.zip, .rar)
  • Executable files (.exe, .app)
Need support for a specific format? Contact us at hello@trypulse.ai

​
Format Detection

Pulse API automatically detects file types based on:
  1. File extension
  2. MIME type
  3. File content analysis
# API automatically detects format
result = client.extract(file_path="document")  # Works even without extension

​
Best Practices by Document Type

Invoices, Statements, Reports
  • Use PDF format when possible
  • Ensure text is selectable (not scanned)
  • Include all pages for context
  • Use schema extraction for structured data
Manuals, Specifications, Diagrams
  • Use extract_figure=True for diagrams
  • Process in smaller chunks if very large
  • Maintain original format for tables
  • Consider HTML output for preservation
Clinical Notes, Lab Reports, Prescriptions
  • Use high-resolution scans (300+ DPI)
  • Process handwritten content carefully
  • Verify extracted data accuracy

​
Troubleshooting

​
Common Issues

Solution: Check that your file extension matches our supported formats
# Correct
client.extract(file_path="document.pdf")

# Incorrect
client.extract(file_path="document.doc")  # Use .docx instead
Solution: Reduce file size or use async processing
# For large files
client.extract(file_path="large_file.pdf", async_mode=True)
Solution: Verify file integrity and re-save if needed
# Check PDF integrity
pdfinfo document.pdf

​
Next Steps

Schema Extraction

Extract structured data from any format

Large Documents

Handle big files efficiently