shield_lock
JinDaGe
← Back to skills
extension
Category: Data & AnalyticsNo API key required

firecrawl 搜索（无需apikey版本）

Search the web and scrape pages using Firecrawl v2 API in keyless mode (no API key required). Use when the user asks to search the web, look up information, research topics, find current news, scrape a URL, or extract web page content. Supports web/news/image search with optional page scraping, and single-page scraping to clean markdown. Triggers: search, look up, find, research, web search, scrape, crawl, firecrawl, extract, read page.

personAuthor: user_38816413hubcommunity
downloadDownload package

Firecrawl Search & Scrape

Keyless web search and page scraping via Firecrawl v2 API. No API key needed. Stdlib-only Python, no pip install. API key resolution: --api-key flag > FIRECRAWL_API_KEY env var > keyless mode.

Search

python scripts/firecrawl_search.py "query" [options]

| Flag | Default | Purpose | |------|---------|---------| | --limit N | 5 | Max results (1-100) | | --sources | web | web, images, news (comma-separated) | | --country | US | ISO country code | | --include | - | Only these domains | | --exclude | - | Exclude these domains | | --categories | - | github, research, pdf | | --scrape | off | Include page markdown in results | | --output | text | text, json, markdown | | --api-key | - | Optional (or set FIRECRAWL_API_KEY env) |

python scripts/firecrawl_search.py "latest AI news" --limit 5
python scripts/firecrawl_search.py "AI breakthroughs" --sources web,news --limit 5
python scripts/firecrawl_search.py "Python tutorial" --scrape --limit 3
python scripts/firecrawl_search.py "react hooks" --include "react.dev,medium.com"
python scripts/firecrawl_search.py "LLM fine-tuning" --categories research
python scripts/firecrawl_search.py "climate policy" --output markdown

Scrape (single page)

python scripts/firecrawl_scrape.py "https://example.com" [options]

| Flag | Default | Purpose | |------|---------|---------| | --formats | markdown | markdown, html, rawHtml, links, screenshot | | --only-main | off | Strip nav/footer, keep main content only | | --output | text | text, json |

python scripts/firecrawl_scrape.py "https://docs.python.org/3/tutorial/" --only-main
python scripts/firecrawl_scrape.py "https://example.com" --formats markdown,links

Research Strategy

For complex topics, run multiple targeted searches instead of one broad query:

  1. Broad scan: --limit 5 to survey the landscape, identify key sources and terminology.
  2. Focused search: use --include to target authoritative domains found in step 1.
  3. Deep read: use firecrawl_scrape.py on the 2-3 most relevant URLs for full content.

Gotchas

  • Uses curl internally (best keyless compatibility); falls back to urllib if unavailable.
  • UTF-8 output handled automatically on Windows (CJK-safe).
  • HTTP 403 = IP blocked in keyless mode > provide --api-key or set FIRECRAWL_API_KEY.
  • HTTP 429 = rate limited > reduce --limit or wait.
  • ~2 credits per 10 results on free tier.