Summary

This project integrates Grok's real-time search capabilities into Claude AI models via the Model Context Protocol (MCP). It addresses AI knowledge cutoffs and enhances fact-checking by providing up-to-date, structured web information directly within Claude's workflow, benefiting users who require current data for research or verification.

How It Works

GrokSearch functions as an MCP server built on FastMCP, acting as a bridge to the Grok API. It intercepts Claude's requests for external information, queries Grok, and returns results in an AI-friendly JSON format containing titles, URLs, and summaries. This approach allows Claude to access current web data, overcoming its training data limitations and improving the accuracy and timeliness of its responses.

Quick Start & Requirements

• Installation: Requires Python 3.10+ and the uv package manager. Installation is performed via claude mcp add-json grok-search --scope user '...', necessitating GROK_API_URL and GROK_API_KEY environment variables.
• Prerequisites: Claude Code or Claude Desktop, and a configured Grok API endpoint. Windows users are strongly recommended to use WSL.
• Verification: Use claude mcp list and prompt Claude with "Please test Grok Search configuration" to validate the setup.
• Documentation: uv installation, GrokSearch GitHub.

Highlighted Details

• Provides an OpenAI-compatible interface for configuration.
• Supports real-time web search and content scraping, with options to target specific platforms like Twitter, Reddit, and GitHub.
• Features tool routing control to disable Claude's native WebSearch/WebFetch and enforce GrokSearch usage.
• Automatically injects local time into queries for improved accuracy on time-sensitive requests.
• Delivers structured JSON output optimized for AI comprehension and citation.

Maintenance & Community

The project is open-source under the MIT license. Community engagement can be tracked via the Star History Chart. No explicit community channels (Discord/Slack) or contributor details are provided in the README.

Licensing & Compatibility

Licensed under the permissive MIT License, allowing for broad use and integration. It is designed for compatibility with Claude via the MCP protocol.

Limitations & Caveats

Windows users are strongly advised to use WSL for optimal setup. The integration requires obtaining and configuring a Grok API endpoint and key. The project focuses on enhancing Claude's capabilities and does not appear to be in alpha or beta status, but specific performance benchmarks beyond the comparison table are not detailed.