Scrape or crawl a web page to clean Markdown for LLMs
crawl4ai.web.scrapeScrape one URL, a web page or a PDF, into clean, LLM-ready Markdown text and/or HTML. Crawl4AI picks the cheapest engine that works (archive, plain fetch, real browser that renders JavaScript), with optional links, media, metadata and tables; proxies for sites that block datacenters.
- 价格
- $0.00015/success按供应商原价
- 成功率
- 51.4%近 30 天 · 1,410,303 次
- 延迟 p50
- 1 秒p95 15 秒
- 最后验证
- 2026年10月5日实测可用
POSThttps://toolswitch.dev/v1/tools/crawl4ai.web.scrape预览- 可用平台额度调用
- 同步返回
- 读取公开数据
参数
Most pages answer in 1-2 s; a page that needs a real browser takes 10-30 s, and a failed page ends within about 55 s. A PDF URL returns the PDF's text as Markdown. Images and other binary files answer 422 binary-body within about a second, free. Blocked (403), unreachable (502) and fleet-busy (503) answers are free; retry a 503 after a few seconds.
这个工具没有参数，详见供应商文档 ↗ 供应商文档
代码示例
将 $TOOLSWITCH_KEY 替换为你在控制台创建的密钥。
curl -X POST 'https://toolswitch.dev/v1/tools/crawl4ai.web.scrape' \
-H "X-Toolswitch-Key: $TOOLSWITCH_KEY" \
-H "Content-Type: application/json" \
--data-raw '{"url":"https://example.com","format":"md"}'import os
import json
import requests
response = requests.request(
"POST",
"https://toolswitch.dev/v1/tools/crawl4ai.web.scrape",
headers={"X-Toolswitch-Key": os.environ["TOOLSWITCH_KEY"]},
json=json.loads("{\"url\":\"https://example.com\",\"format\":\"md\"}"),
)
response.raise_for_status()
print(response.text)const response = await fetch("https://toolswitch.dev/v1/tools/crawl4ai.web.scrape", {
method: "POST",
headers: {
"X-Toolswitch-Key": process.env.TOOLSWITCH_KEY ?? "",
"Content-Type": "application/json",
},
body: JSON.stringify(JSON.parse("{\"url\":\"https://example.com\",\"format\":\"md\"}")),
});
if (!response.ok) throw new Error(`Request failed: ${response.status}`);
console.log(await response.text());网关接口为预览版，地址与鉴权方式可能调整。 示例参数仅为样例数据。 调用按 $0.00015/success 计费，已含服务费。