菜单
Self-hosted Firecrawl — deep web crawling, scraping, and search for AI agents. Converts websites into clean markdown.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Automated research pipeline — find, index, and surface insights from papers, docs, and web sources.
Use this skill for Word documents (.docx) — create, read, edit, format with headings, tables, images, page numbers.
Use this skill for anything involving PDF files — read, create, merge, split, extract text, add watermarks, fill forms, encrypt/decrypt, OCR.
Qorven platform self-knowledge — architecture, capabilities, installable plugins, and self-evolution guide.
Use this skill for PowerPoint presentations — create, read, edit .pptx files with slides, layouts, images, charts.
Create or update Qorven agent skills with structured SKILL.md files, scripts, and references.
| name | firecrawl |
| description | Self-hosted Firecrawl — deep web crawling, scraping, and search for AI agents. Converts websites into clean markdown. |
| when_to_use | When you need to crawl entire websites, extract documentation, scrape JS-rendered pages, or get structured data from complex sites. NOT for single URL reads (use web_fetch) or searching (use web_search). |
| allowed_tools | ["exec","file_write"] |
| context | inline |
Firecrawl converts websites into clean, LLM-ready markdown. It handles JavaScript rendering, follows links, and extracts structured content.
docker run -d --name firecrawl \
-p 3002:3002 \
-e FIRECRAWL_PORT=3002 \
ghcr.io/firecrawl/firecrawl:latest
Add this to your docker-compose.yml:
firecrawl:
image: ghcr.io/firecrawl/firecrawl:latest
ports: ["3002:3002"]
environment:
- FIRECRAWL_PORT=3002
restart: unless-stopped
Set the environment variable so Qorven uses your self-hosted instance:
export FIRECRAWL_URL=http://localhost:3002/v1
| Need | Tool | Cost |
|---|---|---|
| Search the web | web_search | Free (Tavily/DDG) |
| Read one URL | web_fetch | Free (built-in) |
| Deep crawl a site | firecrawl (this plugin) | Free if self-hosted |
| Custom CSS extraction | scrapling plugin | Free |
| JS-rendered pages | firecrawl scrape mode | Free if self-hosted |
Get an API key from https://firecrawl.dev — set CRAWL4AI_API_TOKEN=fc-xxx