codebolt.browser
Browser automation and web interaction tools for controlling web browsers programmatically.
Available Tools
launch
- Launch browser at specified URLclick
- Click at specific coordinatestype
- Type text into active elementscroll_down
- Scroll page downscroll_up
- Scroll page upclose
- Close the browsergetHtml
- Get HTML of current pagegetContent
- Get text content of current pagegetMarkdown
- Get markdown version of current pagescreenshot
- Take screenshot of current pageextractText
- Extract text from current pagebrowserEnter
- Press Enter keygetPdf
- Get PDF version of current page
Sample Usage
// Launch browser at a specific URL
const launchResult = await codeboltMCP.executeTool(
"codebolt.browser",
"launch",
{ url: "https://example.com" }
);
// Click at specific coordinates
const clickResult = await codeboltMCP.executeTool(
"codebolt.browser",
"click",
{ x: 100, y: 200 }
);
// Type text into active element
const typeResult = await codeboltMCP.executeTool(
"codebolt.browser",
"type",
{ text: "Hello World" }
);
// Take a screenshot
const screenshotResult = await codeboltMCP.executeTool(
"codebolt.browser",
"screenshot",
{}
);
// Get page content as markdown
const markdownResult = await codeboltMCP.executeTool(
"codebolt.browser",
"getMarkdown",
{}
);
info
This functionality is similar to the browser API and provides the same capabilities through MCP interface.