Skip to main content

​
Overview

The grok_web_search tool provides focused web content search capabilities, specifically optimized for articles, blogs, and informational pages. It supports both basic results and comprehensive analysis with detailed insights.

​
Parameters

​
query
string
required
The web search query. Must be a non-empty string with a maximum length of 1000 characters.
​
analysis_mode
string
default:"basic"
Analysis depth and format. Must be one of:
  • basic - Returns simple search results with titles, snippets, and URLs
  • comprehensive - Provides detailed analysis with timelines, quotes, multiple perspectives, and context (500+ word analysis)
Default: basicNote: Comprehensive analyses are cached for 30 minutes.
​
max_results
number
default:10
Maximum number of results to return.Constraints:
  • Minimum: 1
  • Maximum: 20
  • Default: 10
​
from_date
string
Optional start date for search in ISO8601 format (YYYY-MM-DD). Limits search to content from this date onwards.Validation:
  • Must match pattern: ^\d{4}-\d{2}-\d{2}$
  • Must be a valid calendar date
  • Must be before or equal to to_date if both are provided
Example: 2025-01-01
​
to_date
string
Optional end date for search in ISO8601 format (YYYY-MM-DD). Limits search to content up to this date.Validation:
  • Must match pattern: ^\d{4}-\d{2}-\d{2}$
  • Must be a valid calendar date
  • Must be after or equal to from_date if both are provided
Example: 2025-12-31

​
Request Examples

{
  "name": "grok_web_search",
  "arguments": {
    "query": "Model Context Protocol MCP",
    "max_results": 5
  }
}

​
Response Format

​
Basic Mode Response

{
  "query": "Model Context Protocol MCP",
  "analysis_mode": "basic",
  "results": [
    {
      "title": "Introduction to Model Context Protocol",
      "snippet": "The Model Context Protocol (MCP) is an open standard that enables seamless integration between AI applications and data sources...",
      "url": "https://modelcontextprotocol.io/introduction",
      "source": "MCP Documentation",
      "published_date": "2024-11-25",
      "author": "Anthropic",
      "citation_url": "https://modelcontextprotocol.io/introduction",
      "citation_index": 0,
      "citation_metadata": {
        "index": 0,
        "url": "https://modelcontextprotocol.io/introduction",
        "domain": "modelcontextprotocol.io",
        "protocol": "https",
        "is_secure": true,
        "path": "/introduction"
      }
    },
    {
      "title": "Building MCP Servers",
      "snippet": "Learn how to build Model Context Protocol servers to expose data and functionality to AI applications...",
      "url": "https://modelcontextprotocol.io/docs/building-servers",
      "source": "MCP Documentation",
      "published_date": "2024-12-01",
      "citation_url": "https://modelcontextprotocol.io/docs/building-servers",
      "citation_index": 1
    }
  ],
  "citations": [
    "https://modelcontextprotocol.io/introduction",
    "https://modelcontextprotocol.io/docs/building-servers"
  ],
  "citation_metadata": [
    {
      "index": 0,
      "url": "https://modelcontextprotocol.io/introduction",
      "domain": "modelcontextprotocol.io",
      "protocol": "https",
      "is_secure": true,
      "path": "/introduction"
    },
    {
      "index": 1,
      "url": "https://modelcontextprotocol.io/docs/building-servers",
      "domain": "modelcontextprotocol.io",
      "protocol": "https",
      "is_secure": true,
      "path": "/docs/building-servers"
    }
  ],
  "summary": "The Model Context Protocol is an open standard for connecting AI applications with data sources, enabling better context sharing and integration.",
  "total_results": 5,
  "search_time": "2025-03-04T12:00:00Z",
  "source": "grok-live-search"
}

​
Comprehensive Mode Response

See grok_search comprehensive response for the full comprehensive analysis structure. The grok_web_search tool returns the same comprehensive format with web-focused content.

​
Error Response

{
  "error": "Search query must be a non-empty string",
  "status": "failed",
  "query": "",
  "search_type": "web",
  "analysis_mode": "basic",
  "timestamp": "2025-03-04T12:00:00Z",
  "request_id": "req_1709560800000_xyz789",
  "from_date": null,
  "to_date": null
}

​
Common Error Messages

  • Search query must be a non-empty string - Query parameter is missing or empty
  • Search query too long (max 1000 characters) - Query exceeds maximum length
  • from_date must be in ISO8601 format (YYYY-MM-DD) - Invalid date format
  • from_date must be before or equal to to_date - Invalid date range
  • API service is not healthy - missing XAI_API_KEY - API key not configured

​
Usage Notes

  • Content Focus: Optimized for articles, blogs, documentation, and general web pages
  • Search Sources: Uses web content sources exclusively (no news or social media)
  • Performance: Basic mode provides faster responses than comprehensive mode
  • Caching: Comprehensive analyses are cached for 30 minutes to improve performance
  • Token Usage: Basic mode uses ~2000 tokens, comprehensive mode uses ~4000 tokens
  • Best For: Technical documentation, tutorials, blog posts, and general information queries
While grok_search supports multiple search types (web, news, general), grok_web_search is specifically optimized for web content:
  • grok_web_search: Fixed to web content sources only
  • grok_search: Configurable search type (web, news, general, twitter)
Use grok_web_search when you specifically want web content without news or social media results.

Build docs developers (and LLMs) love

Get started for freeTalk to us