Crawl4ai

OpenCode plugin for Crawl4AI — gives your AI agent reliable web scraping, deep crawling, LLM-powered extraction, and structured data via the Crawl4AI CLI.

Installation

1. Install the OpenCode plugin

npm install -g @ckanthony/opencode-crawl4ai

Or add it to your opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "plugins": ["@ckanthony/opencode-crawl4ai"]
}

Then install the Crawl4AI CLI globally:

pip install -U crawl4ai
# or
pipx install crawl4ai

After installation, run the browser setup once:

crwl --setup

What it does

This plugin registers the Crawl4AI CLI skill with OpenCode. Once installed, the agent can:

• Scrape any webpage to clean markdown, HTML, or structured data
• Deep crawl entire websites recursively (BFS or DFS)
• LLM Q&A — ask questions about page content using AI
• Structured extraction — extract data matching a JSON schema via LLM
• Content filtering — filter pages by relevance, regex, or cosine similarity
• Parallel crawling — run multiple scrapes in parallel for speed

All output is written to a .crawl4ai/ directory to avoid flooding context.

Docker Server Mode (Optional)

If you prefer to run Crawl4AI as a server instead of using the CLI, you can start it via Docker:

docker pull unclecode/crawl4ai:latest
docker run -d -p 11235:11235 --shm-size=1g unclecode/crawl4ai:latest

Then configure the agent to use http://localhost:11235 as the base URL. The skill covers both CLI and REST API usage.

Links

• Crawl4AI Documentation
• Crawl4AI GitHub
• Crawl4AI CLI Docs
• OpenCode Plugin Docs

License

MIT