Documentation/Document Intelligence

Document Intelligence

Experimental10 credits

Document Intelligence (V2) uses advanced AI models and layout analysis for document processing. Extract text, tables, and structured content from PDFs, images, and Office documents.

Production Recommendation

This is a direct endpoint for development and testing. For production workloads, use the Data Intelligence Pipeline -- it provides structured Data Packages with quality metrics, is async by default, and is covered by Enterprise SLAs.

Overview

Document Intelligence (V2) uses advanced AI models and layout analysis for document processing. Extract text, tables, and structured content from PDFs, images, and Office documents.

Key features:

•Supports PDF, images, DOCX, XLSX, PPTX, HTML
•Layout detection and table extraction
•Chart and seal recognition
•Multiple output formats (Markdown, JSON, HTML, XLSX)
•Performance mode with auto-refinement

API Reference

POSThttps://api.latence.ai/api/v1/document_intelligence/process

Process documents

Request Parameters

Parameter	Type	Default	Description
`file_url`	`string`	—	Public URL to document
`file_base64`	`string`	—	Base64-encoded file data
`filename`	`string`	—	Filename for type detection
`mode`	`string`	—	Processing mode: default or performance
`output_format`	`string`	—	Output: markdown, json, html, xlsx
`max_pages`	`integer`	—	Maximum pages to process
`pipeline_options`	`object`	—	Pipeline configuration

Response Fields

Field	Type	Description
`content`	`string`
`pages_processed`	`number`
`output_format`	`string`
`success`	`boolean`
`usage`	`object`

Response Example

200 OKJSON

{
  "content": "# Document Title\n\nExtracted text content...",
  "pages_processed": 3,
  "output_format": "markdown",
  "success": true,
  "usage": { "credits": 15.0 }
}

Code Examples

from latence import Latence

client = Latence(api_key="YOUR_API_KEY")

# Process a document from URL
result = client.experimental.document_intelligence.process(
    file_url="https://example.com/document.pdf",
    output_format="markdown"  # or "json", "html", "xlsx"
)

print(result.content)  # Extracted text in markdown format
print(f"Pages processed: {result.pages_processed}")

# Or from a local file with performance mode
result = client.experimental.document_intelligence.process(
    file_path="/path/to/document.pdf",
    mode="performance",  # Auto-refinement enabled
    pipeline_options={
        "use_chart_recognition": True,
        "use_seal_recognition": True
    }
)

Explore Tutorials & Notebooks

Deep-dive examples and interactive notebooks in our GitHub repository

View on GitHub

Looking for production-grade processing?

The Data Intelligence Pipeline chains services automatically and returns structured Data Packages.

Pipeline Guide Get API Key