Skip to content

Open-source Python library for AI browser agents, plus Browser Use Cloud: stealth browsers and a hosted agent API.

Tool description

Magnus Müller4212e73d7d46
diff --git a/browser_use/controller/service.py b/browser_use/controller/service.py
index cc28e2d5d..fd5ddb2d5 100644
--- a/browser_use/controller/service.py
+++ b/browser_use/controller/service.py
@@ -328,7 +328,8 @@ class Controller(Generic[Context]):
This tool takes the entire markdown of the page and extracts the query from it.
Set extract_links=True ONLY if your query requires extracting links/URLs from the page.
Only use this for specific queries for information retrieval from the page. Don't use this to get interactive elements - the tool does not see HTML elements, only the markdown.
- Note: Extracting from the same page will yield the same results unless more content is loaded (e.g., through scrolling for dynamic content, or new page is loaded) - so one extraction per page state is sufficient except you change the query.
+ Note: Extracting from the same page will yield the same results unless more content is loaded (e.g., through scrolling for dynamic content, or new page is loaded) - so one extraction per page state is sufficient.
+ If there is no good result of this tool even you see the content is loaded, use the browser state and scrolling to get the information you need.
""",
)
async def extract_structured_data(