菜单
Fetch and extract content from any URL, bypassing bot protection, CAPTCHAs, and geo-blocks using Bright Data's Web Unlocker. Returns clean Markdown by default. For multiple URLs use scrape_batch. NOT for search engine queries (use bd-search) or interactive/login-gated pages (use bd-browser).
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
基于 SOC 职业分类
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/brightdata/brightdata-cursor-plugin --skill bd-scrape命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Process up to 10 URLs or search queries in a single parallel call. Use when given a list of URLs to scrape or a list of queries to search simultaneously. Significantly faster than running them one by one.
Real browser automation for JS-heavy, interactive, or login-gated pages. Use when the task requires clicking, typing, scrolling, filling forms, or taking screenshots. Supports a Puppeteer/Playwright-style flow via MCP tools. Requires Browser zone (PRO tier). Use bd-scrape for static pages that do not need interaction.
Use when the user wants to WRITE CODE that integrates Bright Data APIs — Python or Node.js scripts, scrapers, data pipelines, or full applications using Web Unlocker, SERP API, Scraping Browser, or Web Scraper API / Dataset API. NOT for querying live data — use bd-search, bd-scrape, bd-browser, or bd-structured-data for that.
| name | bd-scrape |
| description | Fetch and extract content from any URL, bypassing bot protection, CAPTCHAs, and geo-blocks using Bright Data's Web Unlocker. Returns clean Markdown by default. For multiple URLs use scrape_batch. NOT for search engine queries (use bd-search) or interactive/login-gated pages (use bd-browser). |
| compatibility | Requires Bright Data MCP server with API_TOKEN and WEB_UNLOCKER_ZONE configured. |
| allowed-tools | mcp(Bright Data:scrape_as_markdown), mcp(Bright Data:scrape_as_html), mcp(Bright Data:scrape_batch) |
| metadata | {"author":"Bright Data"} |
Extract content from: $ARGUMENTS
Single URL — clean readable output (default):
scrape_as_markdown({ url: "<target url>" })
Single URL — raw HTML needed (DOM parsing, link extraction):
scrape_as_html({ url: "<target url>" })
Multiple URLs (up to 10) — batch processing:
scrape_batch({
urls: ["<url1>", "<url2>", "<url3>"]
})
Web Unlocker handles proxy rotation, CAPTCHA solving, and JavaScript rendering automatically. You do not need to configure any of this.
Present extracted content as:
Then the content, following these rules:
[Page Title](url)For batch results, present each URL's content in sequence, labeled with its URL.
If the page content is insufficient, consider:
bd-search to find additional sourcesbd-browser if the page requires login or user interaction to reveal contentIf the tool is unavailable, stop immediately. Do NOT fetch the URL yourself or use built-in fetch tools. Tell the user:
/bd-setup to configure and verify the connection