Perform high-speed web and X (Twitter) searches with structured JSON output and verified citations using the xAI Grok API.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install grok-search
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install grok-search using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Grok Search?
The Grok Search skill enables AI agents to interface directly with the xAI Responses API, bridging the gap between static LLM knowledge and real-time information. By utilizing specialized server-side tools like web_search and x_search, this skill offers a robust alternative to traditional search engines, specifically optimized for the Openclaw Skills ecosystem. It provides deep integration with the X social graph, allowing agents to pull tweets, threads, and user data in a structured format suitable for automated processing.
Beyond simple search, it supports multimodal chat interactions and model management, ensuring developers have full control over the xAI inference lifecycle. Whether you are building research agents or monitoring social trends, this skill provides the necessary plumbing for high-fidelity data retrieval using current models like grok-4-1-fast.
Grok Search Use Cases
- Monitoring real-time discussions and trending topics on X/Twitter.
- Conducting deep-web research with structured citations for LLM grounding.
- Automating social media sentiment analysis by fetching specific handle activities.
- Building multimodal agents that can process both text and images via Grok vision models.
- Verifying facts using the xAI server-side search tools as an alternative to Brave or Google.
How Grok Search Works
- The user or agent triggers a search command through the provided Node.js scripts included in this Openclaw Skills package.
- The skill authenticates with the xAI API using the configured XAI_API_KEY found in environment variables or configuration files.
- A request is sent to the Responses API specifying either a web or X-specific search tool based on the user flags.
- The skill processes the raw API payload, merging and validating citations from response annotations to ensure accuracy.
- It outputs a clean, agent-friendly JSON object containing snippets, URLs, and metadata for the LLM to consume.
Grok Search Setup
To use this skill within the Openclaw Skills collection, ensure you have Node.js installed and an active xAI API key.
Set your environment variable:
export XAI_API_KEY='your_api_key_here'
Run a search query:
node {baseDir}/scripts/grok_search.mjs "current tech trends" --web
List available xAI models:
node {baseDir}/scripts/models.mjs
Grok Search Data Schema & Taxonomy
The skill generates structured JSON output designed for seamless ingestion by LLMs. Below is the primary schema for search results:
| Field | Type | Description |
|---|---|---|
query |
string | The original search terms provided. |
mode |
string | Indicates either web or x search mode. |
results |
array | A list of result objects containing title, url, and snippet. |
citations |
array | Verified source URLs extracted from API annotations. |
For X-specific searches, results additionally include author and posted_at fields to provide social context.
Grok Search Advanced Features
- Server-side filtering for X searches including date ranges, specific handle inclusion, and bot exclusion.
- Multimodal support for image-based chat prompts using Grok vision capabilities.
- Debugging modes like --raw to inspect the underlying Responses API payload on stderr.
- Citation validation that cross-references model output with API metadata for high reliability in Openclaw Skills workflows.
- Flexible output formatting including links-only or text-only modes to reduce token overhead.
SKILL.md
Loading
Related Openclaw Skills
Presearch Search
A privacy-focused, decentralized search skill that provides uncensored web results for AI agents via a distributed node network.
Presearch Decentralized Search
A production-ready decentralized search API for AI agents that provides privacy-first, uncensored web results via a distributed node network.
OpenClaw Team Collaboration Server
A secure, multi-user web interface for OpenClaw featuring zero-knowledge encryption and local network accessibility.
Basic Helper Agent
A versatile personal productivity assistant designed to decompose complex tasks into actionable steps and synthesize multi-source information.
Obul Proxy
Automates x402 payment negotiation for AI agents by proxying upstream requests through Obul.
Stratos Storage
A specialized skill for seamless file interaction with the Stratos Decentralized Storage (SDS) network.
METADATA
Featured*
