Tools¶
9 tools in 4 families.
The names, descriptions and schemas are generated from the server itself, so they are exactly what the model receives in tools/list.
An MCP server that lets an agent search the web and read pages through a real browser. It answers two questions and keeps them apart: where is this written about (queries in, links out) and what does this page say (an address in, its text out). A search never returns the content of a page, and a read never takes a query. Everything else it offers is a way of reading more, or more closely.
Every answer is JSON with an ok flag. A failure carries error, a hint saying what that kind of failure calls for, a type a program branches on — arguments, privateAddress, unanswered, nothingRead, or the kind of a page that failed, such as blocked or robotsDisallowed — and whatever was learned before it; in the protocol it is flagged isError.
| Tool | Family | What it does |
|---|---|---|
playwright_search |
Searching the web | Find where something is written about: queries in, links out — title, address, a short snippet and the engine that offered each. |
playwright_fetch |
Reading pages and sites | Read one page: its address in, its title and readable text out, read by a real browser as a person sees it. |
playwright_fetch_many |
Reading pages and sites | Read several pages in one call: addresses in, a piece of each page's text out, and a page that failed listed beside the others rather than failing the call. |
playwright_fetch_site |
Reading pages and sites | Read a site rather than a page: one address to start from, and the pages of that site are found in its own sitemap, or by following its links when it publish… |
playwright_extract |
Reading a page more closely | Take pieces out of a page instead of its whole text: a CSS selector or an XPath in, exactly one, and the text of each element it matches out — or one attribu… |
playwright_extract_fields |
Reading a page more closely | Take a record of named values out of a page. |
playwright_screenshot |
Reading a page more closely | Look at a page instead of reading it: an address in, a picture of what the browser drew out. |
playwright_perform |
Reading a page more closely | Use a page instead of only reading it: open an address, make a few moves on it, and get the text of the page that results. |
playwright_diagnose |
Diagnostics | Say what state this server is in: its version, whether the browser it reads pages through answers, and which search engines it asks. |