Skip to main content

​
Overview

WraithBytes provides multiple output formats optimized for different use cases. All formats are designed to be token-efficient — delivering maximum information density with minimal token consumption when fed to LLMs.

​
Available Formats

​
Markdown

Clean, pre-processed Markdown that preserves content structure while stripping unnecessary HTML noise. Ideal for LLM consumption. 
{
  "url": "https://example.com/article",
  "options": {
    "markdown": true
  }
}

​
Structured JSON

Typed, structured JSON output from platform-specific parsers. Returns well-defined fields with consistent types. 
{
  "url": "https://www.linkedin.com/in/username",
  "options": {
    "json": true
  }
}

​
HTML

Raw or cleaned HTML for cases where you need the full page source. 
{
  "url": "https://example.com",
  "options": {
    "include_html": "cleaned"
  }
}
Set to "cleaned" for sanitized HTML or "raw" for the original source. Extract all links found on the page. 
{
  "options": {
    "include_links": true
  }
}

​
Image URL Extraction

Extract all image URLs from the page. 
{
  "options": {
    "image_urls": true
  }
}

​
Combining Formats

Options are not mutually exclusive. You can request multiple formats in a single request: 
{
  "url": "https://example.com",
  "options": {
    "markdown": true,
    "include_links": true,
    "image_urls": true
  }
}
This still costs 1 token — output format selection does not affect pricing.

​
Token Efficiency

WraithBytes outputs are specifically designed for AI consumption:
  • Markdown strips boilerplate HTML, ads, and navigation — reducing token count by 60-90% compared to raw HTML
  • Structured JSON provides typed fields that LLMs can reference directly without parsing
  • Custom JSON Schema lets you extract only the fields you need, minimizing payload size
See JSON Schema for custom extraction.