Skip to content

BrowserControl

Tools reference

adityasasidhar/browsercontrol

Tools reference¶

BrowserControl exposes 39 MCP tools organized into seven categories. Every tool that touches the page returns an annotated screenshot plus a textual element map, so the model always has the latest visual context.

Element IDs are fresh on every response

The [1], [2], … numbers you see in a tool's response are only valid until the next screenshot. After any navigation, click, or state change, read the new IDs from the latest screenshot/map. The tools re-screenshot automatically — your agent just needs to look at the response.

Categories¶

Navigation

Navigate the browser: open URLs, history, scroll.
Interaction

Click, type, hover, press keys, wait.
Tabs

Open, switch, close, and list browser tabs.
Forms

Dropdowns, checkboxes, file uploads.
Content

Read the page: text, markdown, JS evaluation, screenshots.
DevTools

Console, network, errors, performance, cookies, viewport.
Recording

Capture Playwright traces and DOM snapshots for debugging.

Quick index¶

Tool	Description
`navigate_to(url)`	Open a URL. Auto-falls back to `127.0.0.1` for `localhost` if the proxy blocks it.
`go_back()`	Browser history back.
`go_forward()`	Browser history forward.
`refresh_page()`	Reload the current page.
`scroll(direction, amount)`	Scroll the viewport.

Interaction¶

Tool	Description
`click(element_id)`	Click by SoM number. Resolves the actual DOM element first, so overlays don't fool it.
`click_at(x, y)`	Click raw coordinates.
`type_text(element_id, text)`	Atomic `element.fill()` — reliable for forms.
`press_key(key)`	Any keyboard key: `Enter`, `Tab`, `Escape`, arrows, etc.
`hover(element_id)`	Hover for tooltips/menus.
`scroll_to_element(element_id)`	Scroll the element into view.
`wait(seconds)`	Sleep — for animations or lazy-loaded content.

Tabs¶

Tool	Description
`create_tab(url=None)`	Open a new tab.
`switch_tab(index)`	Switch by 0-based index.
`close_tab(index)`	Close a tab.
`list_tabs()`	List all tabs with title, URL, active marker.

Forms¶

Tool	Description
`select_option(element_id, option)`	Pick from a `<select>` dropdown.
`check_checkbox(element_id, check=True)`	Toggle a checkbox.
`upload_file(element_id, file_path)`	Native file upload via `set_input_files`.

Content¶

Tool	Description
`get_page_content()`	Whole page as Markdown (script/style stripped).
`get_text(element_id)`	Read text from a specific element.
`get_page_info()`	Current URL + title.
`run_javascript(script)`	Run JS, return serialized result + screenshot.
`screenshot(annotate, full_page)`	Annotated viewport, clean viewport, or clean full-page.

DevTools¶

Tool	Description
`get_console_logs(clear=False)`	Last 50 captured console messages.
`get_network_requests(num_requests, clear=False)`	Method, URL, status, duration (ms).
`get_page_errors()`	Uncaught JS exceptions with stack traces.
`run_in_console(code)`	Eval JS with structured error handling.
`inspect_element(element_id)`	Computed styles, dimensions, attributes.
`get_page_performance()`	TTFB, FCP, DOMContentLoaded, load time.
`get_cookies()`	All cookies in the current context.
`set_cookie(name, value, domain, path)`	Set a cookie (domain auto-inferred).
`delete_cookie(name)`	Delete cookies by name.
`clear_cookies()`	Wipe all cookies.
`set_viewport(width, height)`	Resize the viewport on the fly.

Recording¶

Tool	Description
`start_recording(name="")`	Begin a Playwright trace.
`stop_recording()`	Save the trace.
`take_snapshot(name="")`	Save PNG + HTML + URL triplet.
`list_recordings()`	List recent recordings and snapshots.