Ask Grok is the Grok-focused standalone MCP server in the Ask LLM project, exposing a single ask-grok tool over stdio with an emphasis on exact model selection, an explicit harness, and no fallback whatsoever. It suits developers who already have an xAI API key or Grok CLI, accept metered costs, and want an independent Grok opinion. It is not an official xAI product, requires a secret key, and lacks automatic failover, so weigh cost and data-egress risk before adopting it.
Strongest · Setup experience 13/20 Weakest · Reliability 9/20
Static review · not runReviewed 2026-09-25
Read the FMRS scoring method →Runtime：Node.js 20+
XAI_API_KEY
xAI API key used only for requests to https://api.x.ai; obtain it from the xAI console.
ASK_GROK_HARNESS
Explicit harness: xai-api (default) or grok-cli; no automatic failover between them.
ASK_GROK_MODEL
Exact model ID for the selected harness (API and CLI default grok-4.7), with no fallback.
ASK_GROK_REASONING_EFFORT
Reasoning effort: low, medium, high, or xhigh; default high.
ASK_GROK_MAX_OUTPUT_TOKENS
xAI API output-token ceiling to bound accidental spend; default 16384.
ASK_GROK_TIMEOUT_MS
Timeout for xAI API requests in milliseconds; default 600000.
GMCPT_LOG_LEVEL
Log verbosity: debug, info, warn, error; default warn.
{
"mcpServers": {
"grok": {
"command": "npx",
"args": [
"-y",
"@ask-llm/grok-mcp"
],
"env": {
"XAI_API_KEY": "your-xai-api-key"
}
}
}
}
Shown for Claude Code. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.
{
"servers": {
"grok": {
"command": "npx",
"args": [
"-y",
"@ask-llm/grok-mcp"
],
"env": {
"XAI_API_KEY": "your-xai-api-key"
}
}
}
}
Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).
claude mcp add grok -e XAI_API_KEY=your-xai-api-key -- npx -y @ask-llm/grok-mcp
Run it in a terminal; replace any <…> placeholders with your own values first.
The ask-grok tool should appear in your client's tool list; sending "ask grok to review my last commit" and getting back a Grok review with the exact model attributed proves the connection.
Once connected, you can ask your AI assistant things like:
No matching tools
Listed from the project's documentation, not tested by this site.
Ask Grok is a stdio-based MCP server: the standalone Grok provider package (@ask-llm/grok-mcp, version 0.1.0) from the Ask LLM project. It lets MCP clients such as Claude Code, Cursor, Codex CLI, and Claude Desktop send one-shot prompts to Grok for code review, plan debate, or a second opinion. Two harnesses are supported explicitly: the default xai-api (metered API at https://api.x.ai) and grok-cli (the official Grok CLI). Grok never fails over between harnesses or models: the exact selected-harness model ID is sent unchanged, with grok-4.7 as the default and reasoning effort defaulting to high. Defaults include an output-token ceiling of 16384 and a timeout of 600000 ms. Note: the same repository ships other provider servers (Codex, Claude, Gemini, Antigravity, Ollama) that are not part of this server.
Upstash's official server providing up-to-date third-party library docs for AI coding assistants
AI-powered NPM package analysis MCP server
An independent control layer that gives AI coding agents budgets, brakes, and receipts
GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language
Data synced 2026-09-25 Read the FMRS scoring method