Grok MCP is a secure, self-hosted MCP server that brings Grok-powered real-time web and X search to AI clients with API key management, quotas, usage analytics, and an embedded administration panel.
View on GitHub ↗grok-mcp is an HTTP-only Model Context Protocol (MCP) server that exposes Grok-powered real-time web search, X/Twitter search, and model discovery to MCP clients.
It does not call the official xAI API directly. Instead, it connects to an existing CLIProxyAPI (CPA) deployment. CPA owns the upstream xAI authentication, while grok-mcp provides MCP transport, client API keys, quotas, usage tracking, and an administration panel.
[!IMPORTANT] This project supports Streamable HTTP only. It does not provide a stdio transport or built-in TLS termination.
- grok-mcp must run as a standalone HTTP service, and MCP clients connect to http://:/mcp. - It cannot be configured as a stdio server that an MCP client launches and communicates with over standard input and output. - The service listens for plain HTTP and does not load HTTPS certificates or private keys or perform TLS handshakes. - For an internet-facing deployment, place a trusted reverse proxy such as Nginx, Caddy, Traefik, Kubernetes Ingress, or a cloud load balancer in front of grok-mcp. The proxy should expose HTTPS and forward requests to grok-mcp over internal HTTP.
- Streamable HTTP MCP endpoint at /mcp - Three read-only MCP tools: - grokwebsearch - grokxsearch - groklistmodels - Selectable CPA upstream protocol: OpenAI Responses, OpenAI Chat Completions, or Anthropic Messages - MCP progress notifications for upstream search rounds - Per-user client API keys with enable/disable controls - Tier-based RPM and monthly successful-call quotas - Valid-forwarded-IP-triggered protection for /mcp and panel authentication - SQLi
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/MapleMapleCat--Grok_Search_Mcp.html)
Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zer
vshulcz/deja-vuYour agents already solved this. deja finds it — retroactive search and MCP recall over the session logs your coding agents already wrote. Cross-agent, SSH-portable, one zero-dep binary, fully local.
olucurious/watchtowerSelf-hosted error tracking. A drop-in for Sentry and AppSignal SDKs: one Go binary and Postgres, with Slack, Linear, email and an MCP server for coding agents.
KnifeLemon/DroidlineAndroid automation from code. Test your apps on real phones, script the tasks you repeat, run a shelf of phones, or let an AI agent use one, from Python, Node.js, the CLI, HTTP or Appium clients. No A
enowdev/succubusCross-agent coordination for AI coding agents. One daemon, one database — so multiple agents in one repo can see each other's plan, tasks, and file claims.
AsWali/CallBoardA shared to-do list for you and your AI coding agents
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.