Seleniumboot MCP — Selenium Browser Automation

Control a real Chrome/Firefox browser directly from GitHub Copilot and Claude Code using the Model Context Protocol. 82 tools for navigation, interaction, assertions, codegen, and more — no ChromeDriver setup needed.

What it does

Installing this extension:

1. Auto-registers the seleniumboot-mcp server with GitHub Copilot (via contributes.mcpServers)
2. Writes .mcp.json into your open workspace so Claude Code picks it up automatically
3. Prompts you to install seleniumboot-mcp via pip if it isn't already

Once active, your AI assistant can control a real browser just by asking.

Requirements

• Python 3.10+ with pip
• Google Chrome (ChromeDriver is managed automatically — nothing to download)
• VS Code 1.99+ with GitHub Copilot or Claude Code

Quick start

1. Install the Python package

pip install seleniumboot-mcp

If the package isn't installed, the extension shows a notification on startup with an Install via pip button.

2. Approve the MCP server

When you open a project, Claude Code will detect the .mcp.json written by this extension and ask:

"Allow MCP server seleniumboot?" → click Allow

For GitHub Copilot, the server is registered automatically via contributes.mcpServers — no approval needed.

3. Start automating

In Copilot Chat (Agent mode) or a Claude Code chat:

Go to https://example.com and take a screenshot

Fill the login form at https://myapp.com with admin / password and click Login

Assert the dashboard heading is visible, then generate a Java TestNG test for the session

Chrome opens automatically on the first browser command — no explicit "start browser" step required.

Commands

Open the Command Palette (Ctrl+Shift+P) to access:

Available tools (82)

Browser

navigate, take_screenshot, get_page_source, execute_script, go_back, go_forward, refresh, scroll_to_top, scroll_to_bottom, scroll_by, get_page_title, get_current_url, open_new_tab, close_current_tab, switch_to_window, list_windows, close_browser, emulate_device, get_console_logs, get_cookies, set_cookie, delete_cookie, delete_all_cookies, get_local_storage, set_local_storage, get_session_storage, set_session_storage, wait_for_network_idle, inspect_page, get_network_logs, mock_response, clear_mock_responses, compare_screenshot, check_accessibility

Elements

find_element, find_elements, click, type_text, get_text, get_attribute, select_option, hover, double_click, right_click, drag_and_drop, is_displayed, is_enabled, wait_for_element, scroll_to_element, clear_field, send_keys, upload_file, accept_alert, dismiss_alert, get_alert_text, type_in_alert, switch_to_frame, switch_to_default_content, find_shadow_element, get_table_data, fill_form, get_healed_locators, clear_healed_locators

Assertions

assert_title, assert_url, assert_text, assert_element_visible, assert_element_not_visible, assert_attribute, assert_page_contains, assert_element_count

Code generation

generate_java_testng, generate_java_junit5, generate_java_page_object, generate_gherkin, generate_python_test, generate_csharp_nunit, generate_github_actions, generate_playwright_hints, get_session_log, clear_session_log

Example: generate a Java test

Go to https://myapp.com/login
Enter "admin" in the username field and "password" in the password field
Click the Login button
Assert the text "Dashboard" is visible
Generate a Java TestNG test class for this session

Claude or Copilot records every step and outputs a ready-to-paste test:

public class LoginTest {
    @BeforeMethod public void setUp() { driver = new ChromeDriver(); }
    @AfterMethod  public void tearDown() { driver.quit(); }

    @Test
    public void recordedFlowTest() {
        driver.get("https://myapp.com/login");
        wait.until(visibilityOfElementLocated(By.cssSelector("#username"))).sendKeys("admin");
        driver.findElement(By.cssSelector("#password")).sendKeys("password");
        wait.until(elementToBeClickable(By.cssSelector("button[type='submit']"))).click();
        wait.until(visibilityOfElementLocated(By.xpath("//*[contains(text(),'Dashboard')]")));
    }
}

Troubleshooting

MCP server not connecting in Claude Code

• Open the Command Palette → Seleniumboot MCP: Check installation status
• Ensure a .mcp.json file exists in your project root (the extension creates it automatically)
• Do Developer: Reload Window and open a fresh Claude Code chat

"seleniumboot-mcp not found" error

• Run pip install seleniumboot-mcp in a terminal
• Make sure the Python bin directory is in your PATH

Chrome doesn't open

• Chrome must be installed; ChromeDriver is handled automatically by Selenium Manager
• For headless environments use: "Start browser in headless mode"

Links

• GitHub
• PyPI package
• Demo video
• Report an issue