| name | firecrawl-cli |
| description | Use when the user wants to scrape websites, crawl or map URLs, search the web, extract structured data from web pages, or run AI web agents using the Firecrawl API from the command line. |
| author | alexgorbatchev |
Firecrawl CLI Commands
Use the firecrawl CLI to interact with the Firecrawl API. Note that this is not the official Firecrawl CLI, but rather a small Golang implementation: https://github.com/alexgorbatchev/firecrawl-cli. It supports mapping, scraping, searching, and agent tasks.
Core Capabilities
The firecrawl CLI provides four main commands:
scrape: Scrape a single page and get its content. Extracts structured data, markdown, or HTML.
map: Discover and map all URLs on a website quickly.
search: Search the web and optionally scrape the results.
agent: Search and gather data from the web using natural language prompts.
Important:
- DO NOT set or override the
FIRECRAWL_API_URL environment variable (e.g., export FIRECRAWL_API_URL="https://api.firecrawl.dev"). The system is intentionally configured to use a local private server. Rely on the default environment.
- You can output structured JSON using the
--json flag or format flags.
Quick Examples
-
Scrape to Markdown:
firecrawl scrape "https://example.com" --format markdown --output example.md
-
Scrape Structured Data:
firecrawl scrape "https://example.com" --schema '{"type":"object","properties":{"title":{"type":"string"}}}' --json
-
Map URLs:
firecrawl map "https://example.com" --limit 50 --output urls.json
-
Search and Scrape:
firecrawl search "latest AI news" --scrape --limit 5 --json
-
Run Web Agent:
firecrawl agent "Find the pricing of Mendable and save it" --wait --json
Reference Documentation
For a full list of commands, flags, and advanced options, consult the reference documentation:
- See
references/firecrawl-cli.md for detailed usage instructions and flags.