The Scrape category fetches web pages and returns clean content your agent can
actually use. Discover a tool with find_tools, then call it with type + params.
Providers
| Provider | Operation | Best for |
|---|
| Firecrawl | scrape | Any URL → clean markdown, with caching and proxy options |
| Jina Reader | readUrlGet / readUrlPost | Fast URL → markdown via a simple request |
| BrightData | unlock | Last-resort fallback for Cloudflare / challenge-gated pages (raw HTML) |
- Start with Firecrawl for most pages — it returns structured markdown and
handles PDFs.
- Use Jina Reader for lightweight, fast reads.
- Fall back to BrightData Web Unlocker only when a page is protected by
anti-bot challenges; it returns raw HTML, not markdown.
Example
find_tools("scrape https://example.com main content")
describe_tool("Firecrawl/scrape")
execute_tool("Firecrawl/scrape", { url: "https://example.com", formats: ["markdown"] })
Firecrawl supports caching (maxAge) and onlyMainContent to strip nav,
footers, and boilerplate. Check describe_tool for the full parameter set and
per-call cost.
