Add the BrowserStack MCP server to Gemini CLI

gemini mcp add -e BROWSERSTACK_USERNAME=YOUR_VALUE -e BROWSERSTACK_ACCESS_KEY=YOUR_VALUE browserstack npx -y @browserstack/mcp-server

Scopes, env vars, OAuth and limits: Gemini CLI MCP config guide

Run /mcp auth <name> inside Gemini CLI to sign in to an OAuth server. BrowserStack runs locally and needs BROWSERSTACK_USERNAME, BROWSERSTACK_ACCESS_KEY. Replace the placeholder values above with your own before starting Gemini CLI.

Ask Gemini CLI for something BrowserStack can do. Run Playwright and Selenium tests on real devices and browsers, debug failures, scan for accessibility issues and manage test cases. Approve the first tool call and check the result before allowing write actions automatically.

• · In settings.json, streamable HTTP servers use httpUrl, SSE servers use url, local servers use command.
• · includeTools and excludeTools limit which tools the model sees.

How do I add the BrowserStack MCP server to Gemini CLI?

Run: gemini mcp add -e BROWSERSTACK_USERNAME=YOUR_VALUE -e BROWSERSTACK_ACCESS_KEY=YOUR_VALUE browserstack npx -y @browserstack/mcp-server. Run /mcp auth <name> inside Gemini CLI to sign in to an OAuth server.

Is this the official BrowserStack MCP server?

Yes. It is published by BrowserStack as io.github.browserstack/mcp-server in the official MCP registry.

Does the BrowserStack MCP server need an API key?

Yes. It runs locally and requires BROWSERSTACK_USERNAME, BROWSERSTACK_ACCESS_KEY.