data-ai MCP Server
Bring Your Own Browser — let your AI agent use the Chrome you already have open
Discovered via github-seeds:mcp-hot and last synced 3mo ago.
Install instructions not detected yet
Check the source repository for the latest setup steps.
Open a page, scroll through, read all text
hours
Stop recording, export JSON or HAR
Scroll to top/bottom, an element, or a Y coordinate
yes
What it does
Download all images from a page
Export cookies for `curl` / scripts
Run JavaScript on the page (off by default)
Pull `<table>` elements as JSON
$$
Screenshot → saved to disk
Send a keyboard key (Enter, Escape, F5, ArrowDown, ...)
Start a stateful request-interception session.
Drag the mouse from one point to another (linear interpolation).
Get raw HTML of an element or the whole page
Close a tab by tabId
Stop a `browser_intercept_start` session and return hit stats.
Snapshot console.log / warn / error
Hover the mouse over an element to trigger tooltips/menus
Wait for an element to appear
Page Web Vitals + navigation timing.
URL on the blocklist
Headless Puppeteer
Choose an option in a native `<select>`
List all open tabs
Upload local files to `<input type="file">`.
Extension lost connection
Type into an input (optionally press Enter)
Go back one step in browser history
Save current page as PDF (default `~/.byob/pdfs/`).
Cause
Click a button or link
Write a cookie via `chrome.cookies.set` (CHIPS-aware).
Emulate mobile/tablet viewport / DPR / touch / UA.
Go forward one step in browser history
Read `localStorage` / `sessionStorage` for an origin.
DevTools open on that tab
Switch to a tab
Auto
Same, returns clean markdown (no nav/ads)
Open a URL in a new or existing tab
Start recording HTTP + WebSocket traffic