Antigravity Browser Operator
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| browser_list_profilesB | Lists all Google Chrome user profiles detected on this computer (e.g. Personal, Work, Clean, Hashan), showing profile folder names ("Default", "Profile 13"), profile display names, emails, whether the Browser Operator extension is installed, and real-time WebSocket connection status. |
| browser_launch_profileA | Launches a specific Google Chrome user profile window directly by name, folder directory (e.g. "Profile 13", "Default"), or email, optionally navigating to an initial URL. Bypasses the Chrome profile picker dialog entirely. |
| browser_list_tabsA | Lists all open tabs across all active Chrome windows and profiles, returning tab ID, title, URL, active status, and profile name. |
| browser_select_tabB | Switches to / focuses a specific tab in the user’s Chrome browser by tab ID. |
| browser_new_tabB | Opens a new tab in the user’s Chrome browser with a specified URL, optionally in a designated profile. |
| browser_close_tabB | Closes a specific tab in the user’s Chrome browser by tab ID. |
| browser_batch_actionsA | Executes multiple sequential browser actions (click, type, wait, scroll, select_option, set_checked, navigate) in a single turn without round-trip LLM pauses. Drastically speeds up multi-step tasks, form filling, and searching. |
| browser_navigateA | Navigates the specified or active tab in the user’s Chrome browser to a URL and waits for page load. |
| browser_get_domB | Extracts the page title, URL, token-dense compact DOM representation, indexed interactive elements with refs, and clean text of the tab. |
| browser_clickC | Clicks an element on the active or specified tab using an indexed ref, CSS selector, or visible text. |
| browser_typeC | Types text into an input or textarea element on the active or specified tab. |
| browser_select_optionB | Selects an option from a dropdown or listbox on the active or specified tab. |
| browser_set_checkedB | Sets the checked state of a checkbox, radio button, or switch toggle. |
| browser_scrollC | Scrolls the active or specified tab up or down. |
| browser_screenshotB | Captures a visible screenshot of the active or specified tab in Chrome and saves it as an image on disk. |
| browser_wait_forB | Waits for an element to appear, become visible, clickable, hidden, or waits for network idle (pending API requests to finish). |
| browser_get_console_logsB | Retrieves recent console logs, warnings, runtime errors, and unhandled promise rejections from the page. |
| browser_get_network_logsA | Retrieves recent network activity (Fetch, XHR, resource requests, status codes, failed requests, and response times). |
| browser_evaluateB | Executes arbitrary JavaScript in the context of the active or specified tab and returns the result. |
| browser_upload_fileB | Uploads a local file to a file input on the active or specified tab completely autonomously without opening OS file picker. |
| browser_handle_dialogA | Handles, accepts, or dismisses browser native JavaScript modal dialogs (alerts, confirms, prompts, and "Leave site?" beforeunload dialogs) using Chrome DevTools Protocol. |
| browser_save_recipeA | Saves a successful multi-step workflow/macro into the agent’s muscle memory. Stored recipes can be recalled and executed in seconds for repetitive site tasks. |
| browser_find_recipeA | Searches the agent’s muscle memory for existing recipes on a domain or task. Returns stored action sequences ready to run via browser_batch_actions. |
| browser_learn_site_quirkB | Records a site-specific quirk, pitfall, selector tip, or bypass rule into the long-term knowledge library. |
| browser_get_site_memoryA | Retrieves all learned knowledge, known quirks, and saved recipes for a specific domain or the active tab. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 25 tools
Most tools map to a single distinct browser action (click, type, scroll, navigate, screenshot), so selection is usually unambiguous. The main fuzziness is in the memory group — save_recipe, learn_site_quirk, find_recipe, and get_site_memory overlap in that all store/retrieve per-domain knowledge — and batch_actions partially duplicates the granular action tools, though descriptions clarify the distinction.
Every tool carries the `browser_` prefix and uses snake_case, giving a predictable namespace. Minor deviations exist where the second token is a noun rather than a verb (browser_new_tab, browser_batch_actions), but overall the verb_noun pattern holds well.
At 25 tools this sits at the heavy end of the scale for a browser automation server. Core actions are justified, but the recipe/quirks/memory trio plus logs, profiles, and dialogs push the surface beyond a tight, well-earned set.
Coverage is broad: navigation, tab lifecycle, profiles, interaction, dialogs, uploads, screenshots, DOM extraction, JS eval, and diagnostics. Gaps remain for keyboard key presses (Enter/Tab/Escape), hover and drag-and-drop, history back/forward, and iframe traversal, which agents will hit on real sites.