xAI / Grok Skill for OpenClaw

Full xAI platform skill. Chat, vision, image generation, video generation, text-to-speech, Responses API, Collections/RAG, and batch processing with Grok models.

What it does

• Chat with Grok - text conversations with Grok-4, Grok-4.20, Grok-3, Grok-3-mini, Grok-3-fast
• Vision analysis - analyze images, screenshots, diagrams, OCR documents
• Image generation - text-to-image with grok-imagine-image and grok-imagine-image-pro
• Image editing - edit images with natural language, up to 3 input images
• Video generation - text/image/video to video, up to 10 sec at 720p
• Text-to-speech - 5 voices, 20 languages, expression tags, WebSocket streaming
• Responses API - server-side tools (web_search, x_search, code_interpreter, collections_search)
• Collections/RAG - upload documents, create collections, keyword/semantic/hybrid search
• Reasoning mode - step-by-step chain-of-thought with dedicated reasoning models
• Code generation - write, review, and translate code (dedicated grok-code-fast-1 model)
• Model comparison - run the same prompt through multiple models
• Function calling - structured tool use via OpenAI-compatible format
• Batch API - bulk processing for thousands of requests
• Prompt caching - up to 90% input cost reduction for repeated prefixes

For X/Twitter search, use the search-x skill instead.

Quick start

Install the skill

git clone https://github.com/mvanhorn/clawdbot-skill-xai.git ~/.openclaw/skills/xai

Set up your API key

Get a key from console.x.ai, then:

export XAI_API_KEY="xai-YOUR-KEY"

Example usage

• "Ask Grok about the state of AI in 2026"
• "Grok, analyze this screenshot" (with image)
• "Generate an image of a sunset over mountains"
• "Create a 5-second video of ocean waves"
• "Read this text aloud with the eve voice"
• "Search the web and summarize recent AI news" (Responses API)
• "Walk me through this math problem step by step"
• "Have Grok write a Python web scraper"
• "Compare Grok-3 and Grok-4 on this question"

Available models

Text / Chat

| Model | Context | Notes | |-------|---------|-------| | grok-4.20-multi-agent-beta-0309 | 2M | Multi-agent orchestration | | grok-4.20-beta-0309-reasoning | 2M | Extended reasoning | | grok-4 | 131K | Most capable general model | | grok-4.20 | 131K | Latest general release | | grok-4-1-fast-reasoning | 2M | Fast reasoning, lower cost | | grok-3 | 131K | General-purpose default | | grok-3-mini | 131K | Fast, reasoning effort control | | grok-3-fast | 131K | Lowest latency | | grok-code-fast-1 | 256K | Dedicated code model | | grok-2-vision-1212 | 32K | Vision-capable |

Image / Video

| Model | Price | Notes | |-------|-------|-------| | grok-imagine-image-pro | $0.07/img | Highest quality | | grok-imagine-image | $0.02/img | Fast, cost-effective | | grok-imagine-video | $0.050/sec | Up to 10 sec at 720p |

CLI usage

# Chat
node scripts/chat.js "What is quantum computing?"
node scripts/chat.js --model grok-4 "Explain monads"
node scripts/chat.js --model grok-code-fast-1 "Write an async web crawler in Python"

# Vision
node scripts/chat.js --image screenshot.png "What UI issues do you see?"

# Reasoning
node scripts/chat.js --model grok-3-mini --system "Think step by step." "Prove sqrt(2) is irrational"

# Streaming
node scripts/chat.js --stream "Write a short story"

# List models
node scripts/models.js

API endpoints

| Endpoint | Method | Description | |----------|--------|-------------| | /v1/chat/completions | POST | OpenAI-compatible chat | | /v1/responses | POST | Responses API (recommended) | | /v1/images/generations | POST | Image generation | | /v1/images/edits | POST | Image editing | | /v1/tts | POST | Text-to-speech | | WSS /v1/tts | WS | Streaming TTS | | /v1/files | POST | File upload | | /v1/models | GET | List models |

Base URL: https://api.x.ai Management API: https://management-api.x.ai (collections, document search)

How it works

Wraps the xAI API platform. Chat uses /v1/chat/completions (OpenAI-compatible). The Responses API (/v1/responses) adds server-side tools like web_search, x_search, and code_interpreter. Image generation, video generation, and TTS each have dedicated endpoints. Collections/RAG uses the management API for document upload and search.

• API docs: https://docs.x.ai/api

License

MIT