Browser Jet Pilot
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HOST | No | Bind address | localhost |
| PORT | No | Port for HTTP transport | 3100 (if HTTP mode) |
| LAUNCH | No | Launch new browser instead of connecting | false |
| API_KEY | No | Optional API key to secure HTTP transport | |
| CDP_URL | No | CDP endpoint | http://localhost:9222 |
| BROWSER_WIDTH | No | Viewport width | 1280 |
| BROWSER_HEIGHT | No | Viewport height | 720 |
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 | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| browser_startA | Connect to the browser (or launch a new one). Must be called before any other browser tool. Reuses existing session if still alive. |
| browser_endA | Close the current browser session and release resources. |
| browser_navigateA | Navigate the browser to a URL. Waits for the page to reach domcontentloaded state. |
| browser_get_infoA | Get current page metadata: URL, title, viewport size. |
| browser_screenshotA | Take a screenshot of the current page. Returns a base64-encoded PNG image. |
| browser_clickC | Click an element on the page identified by a CSS selector. |
| browser_fillA | Clear and fill an input field or textarea. Does not trigger change events; use browser_type for character-by-character input. |
| browser_typeA | Type text character by character into a focused element. Triggers keydown/keyup events. Clicks the element first to focus. |
| browser_selectB | Select an option in a dropdown element. |
| browser_hoverB | Hover the mouse over an element. |
| browser_scrollB | Scroll the page or a specific element. |
| browser_evaluateB | Execute JavaScript code in the browser page. The code runs in the page context. Return a JSON-serializable value. |
| browser_get_contentB | Extract text content or HTML from the page. Useful for reading what is currently displayed. |
| browser_disable_shadersA | Inject a script that blocks WebGL, throttles requestAnimationFrame, and freezes CSS animations/transitions. Use on heavy shader-rendered pages (Three.js, WebGL dashboards) to make them readable without GPU strain. Call BEFORE navigating to the target page for best results. |
| browser_restore_shadersA | Restore WebGL, requestAnimationFrame, and CSS animations that were disabled by browser_disable_shaders. Removes the injected style element and restores original browser functions. |
| browser_wait_forB | Wait for a condition before proceeding. Waits for an element to appear, be visible, or be hidden. |
| browser_new_tabB | Open a new browser tab and switch to it. |
| browser_list_tabsB | List all open tabs in the current browser session. |
| browser_switch_tabC | Switch to a different tab by index. |
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 19 tools
Each tool targets a distinct action or purpose: clicking, filling, typing, evaluating scripts, managing tabs, etc. The only potential overlap is between 'browser_fill' and 'browser_type', but their descriptions clearly differentiate them (bulk fill vs. character-by-character typing with events). No confusion between tools.
All tools follow a consistent 'browser_verb' or 'browser_verb_noun' pattern in snake_case (e.g., browser_click, browser_list_tabs, browser_disable_shaders). No mixing of conventions or verb styles.
With 19 tools, the server covers a wide range of browser automation tasks without being bloated. Each tool serves a clear purpose, from basic navigation and interaction to advanced features like shader control. The count feels appropriate for a comprehensive browser automation MCP server.
The tool set covers core browser operations: navigation, clicks, typing, filling, scrolling, select, tab management, screenshots, JavaScript execution, and waiting. Minor gaps exist such as file upload, cookie management, or form submission utilities, but these can often be worked around via JavaScript or clicking. Overall, it's robust for typical use cases.