Enables LLMs like Claude to navigate the web through Puppeteer-based tools and Steel. Based on the Web Voyager framework, it provides tools for all the standard web actions click clicking/scrolling/typing/etc and taking screenshots.
暂无评论. 成为第一个评论的人！
登录以参与讨论
Navigate to any URL in the browser. Inputs: url (string, required): URL to navigate to (e.g. 'https://example.com').
Perform a Google search by navigating to 'https://www.google.com/search?q=encodedQuery'. Inputs: query (string, required): Text to search for on Google.
Click elements on the page using numbered labels. Inputs: label (number, required): The label number of the element to click.
Type text into input fields using numbered labels. Inputs: label (number, required): The label number of the input field, text (string, required): Text to type into the field, replaceText (boolean, optional): If true, replaces any existing text in the field.
Scroll down the page. Inputs: pixels (integer, optional): Number of pixels to scroll down. If not specified, scrolls by one full page.
Scroll up the page. Inputs: pixels (integer, optional): Number of pixels to scroll up. If not specified, scrolls by one full page.
Navigate to the previous page in browser history. No inputs required.
Wait for up to 10 seconds, useful for pages that load slowly or need more time for dynamic content to appear. Inputs: seconds (number, required): Number of seconds to wait (0 to 10).
Capture the current page without bounding boxes or highlights and store it as a resource. Inputs: resourceName (string, optional): Name to store the screenshot under (e.g. 'before_login'). If omitted, a generic name is generated automatically.
