BrowserStack MCP Server

Official

BrowserStack's official MCP server for natural language control of the test platform.

How to install ↓ Compare GitHub source repository ↗

Category: Dev Tools #355 of 438
Stars: ★ 154 Very popular
Transport: stdio (local process)
Runtime: Node.js 18+
Credentials: API key / credential required
License: AGPL-3.0
Last commit: 3 days ago
Tools: 44

40FMRS · D

The BrowserStack MCP server is an official server that offers a comprehensive set of tools for integrating the testing platform into MCP clients. It supports manual and automated testing, accessibility and visual testing, and AI-driven debugging. Installation requires BrowserStack credentials, and some tools are not available in Remote MCP. Overall, it is a powerful integration for teams using BrowserStack.

Strongest · Security and permissions 10/20 Weakest · Reliability 6/20

Reliability

6/20

Security and permissions

10/20

Maintenance

8/20

Documentation

8/20

Setup experience

8/20

Why each score

Reliability 6/20: Evidence insufficient: only static review of README and manifest; no executed evidence to confirm actual reliability. README claims 44 tools with detailed names, but no tests or CI evidence to verify behavior. No blatant contradictions found, but due to lack of executed evidence and many tools with thin error handling, score 6 (below 12 static cap).
Security and permissions 10/20: Credentials passed via environment variables (BROWSERSTACK_ACCESS_KEY) with no hardcoded secrets. Tools like fetchRCA and prepareSelfHealingPlan explicitly state they only suggest fixes and never auto-apply, indicating good permission boundaries. However, side effects like launching sessions lack explicit user confirmation, and data-flow disclosure is limited. No red-line issues; permissions and confirmation incomplete, score 10.
Maintenance 8/20: Repository is active (148 stars, 39 open issues), npm version 1.2.34 vs manifest 1.0.0 shows ongoing updates. README has contributing guidelines and support channels, license AGPL-3.0. But no dependency update policy or security-response channel; versioning gap, score 8.
Documentation 8/20: README provides detailed installation, configuration, all tools with example prompts, and limitations (local vs remote). Missing thorough parameter details, limits (rate/cost), and troubleshooting guide. Many tools but documentation is mainly list of names with examples, lacking depth, score 8.
Setup experience 8/20: Clear installation steps with config examples for VSCode, Cursor, Claude Desktop, plus one-click links. Setup requires API keys and Node >=18, but no mention of platform compatibility beyond these. Config examples are concrete, but no automated tests or CI verification; below static cap 15, score 8.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesUses the networkControls a browser

Best for

QA teams using BrowserStack
Developers who want to integrate testing into AI workflows
Teams needing accessibility compliance
Automated test maintenance

Not for

Teams not using BrowserStack
Users needing local testing support (Remote MCP cannot access localhost)
Users expecting fully automated file modifications (tools only suggest fixes)

Required permissions

Requires BrowserStack username and access key
Can create and modify Test Management projects, test cases, and runs
Can start live sessions and run automated tests
Can access test reports and logs
Some tools require access to local file system (e.g., uploading files)

Risks and side effects

Credentials may be logged by tools; handle them securely
AI might produce unexpected actions; test in a safe environment
Some tools are unavailable in Remote MCP, behavior may differ
Test execution may incur costs
Dependency on network and BrowserStack service availability

Setup

Before you start

Runtime：Node.js 18+

BROWSERSTACK_USERNAME required Your BrowserStack account username, available from Account Settings on browserstack.com.

BROWSERSTACK_ACCESS_KEY requiredsecret Your BrowserStack access key (a secret credential), available from Account Settings on browserstack.com.

Ensure Node.js >= 18.0 (v22.15.0 LTS recommended).
Sign up for a BrowserStack account and get your username and access key.
Add the JSON configuration to your client's config file, or use one-click setup for VS Code/Cursor.
Start the server and authorize.

claude_desktop_config.json

{
  "mcpServers": {
    "browserstack": {
      "command": "npx",
      "args": ["-y", "@browserstack/mcp-server@latest"],
      "env": {
        "BROWSERSTACK_USERNAME": "<username>",
        "BROWSERSTACK_ACCESS_KEY": "<access_key>"
      }
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "browserstack": {
      "command": "npx",
      "args": [
        "-y",
        "@browserstack/mcp-server@latest"
      ],
      "env": {
        "BROWSERSTACK_USERNAME": "<username>",
        "BROWSERSTACK_ACCESS_KEY": "<access_key>"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add browserstack -e 'BROWSERSTACK_USERNAME=<username>' -e 'BROWSERSTACK_ACCESS_KEY=<access_key>' -- npx -y @browserstack/mcp-server@latest

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After configuring the client config, restart your MCP client and confirm BrowserStack tools (e.g., createProjectOrFolder, runBrowserLiveSession) appear in the tool list; then try a prompt like "open browserstack.com on latest Chrome" to confirm a Live session starts.

Troubleshooting

Check Node version is >= 18.0
Verify BrowserStack username and access key are correct
Ensure you have an appropriate plan
Check network connectivity and firewall settings
For local testing, use local MCP server instead of Remote
Open issues on GitHub or contact BrowserStack support

Things to try

Once connected, you can ask your AI assistant things like:

Open my app on an iPhone 15 Pro Max and help me debug a crash on Android 14
Open the website hosted on localhost:3001 in the latest version of Edge
Create a test run for the Login tests in the Test Management project named My Demo Project
Run an accessibility scan on www.bstackdemo.com and identify issues

Tools 44

createProjectOrFolder writes

Create a Test Management project and/or folders, returns project/folder ID and link.

createTestCase writes

Add a manual test case under a specific project/folder.

updateTestCase writes

Update an existing test case with any subset of fields.

listTestCases read-only

List test cases in a project, optionally scoped to a folder.

listFolders read-only

List folders in a Test Management project with IDs, names, and case counts.

listTestCaseTemplates read-only

List available test case templates for use with createTestCase.

createTestRun writes

Create a test run (suite) for selected test cases.

listTestRuns read-only

List test runs for a project, filter by dates, assignee, or state.

Show 36 more tools

updateTestRun writes

Update a test run's name/state and/or add test cases.

addTestResult writes

Add a manual execution result (passed/failed/blocked/skipped) for a test case.

createTestCasesFromFile writes

Generate test cases in bulk from an uploaded file using AI (not available in Remote MCP).

listTestPlans read-only

List test plans in a project.

getTestPlan read-only

Get details of a test plan, including linked runs and status.

listSubTestPlans read-only

List sub-test-plans under a parent test plan.

getSubTestPlan read-only

Get details of a sub-test-plan under a parent plan.

setupBrowserStackAutomateTests writes

Set up BrowserStack SDK to run web tests.

fetchAutomationScreenshots read-only

Fetch screenshots captured during an automation session.

getFailureLogs read-only

Retrieve error logs for automation sessions.

fetchBuildInsights read-only

Get build insights and quality-gate results.

runAppLiveSession writes

Start a manual app testing session on a real device.

runBrowserLiveSession writes

Start a Live session for website testing.

takeAppScreenshot read-only

Launch an app and capture a screenshot to verify launch.

runAppTestsOnBrowserStack writes

Run existing native mobile test suites (Espresso/XCUITest).

setupBrowserStackAppAutomateTests writes

Set up BrowserStack App Automate SDK for Appium mobile testing.

accessibilityExpert read-only

Ask the A11y Expert about WCAG and usability best practices.

startAccessibilityScan writes

Start a web accessibility scan and get a report path.

createAccessibilityAuthConfig writes

Create an authentication config for accessibility scans.

getAccessibilityAuthConfig read-only

Get an existing accessibility auth config by ID.

fetchAccessibilityIssues read-only

Fetch accessibility issues from a completed scan.

percyVisualTestIntegrationAgent writes

Integrate Percy visual testing into a new project.

expandPercyVisualTesting writes

Set up or expand Percy visual testing for existing projects.

addPercySnapshotCommands writes

Add Percy snapshot commands to test files (not available in Remote MCP).

listTestFiles read-only

List all test files in given directories (not available in Remote MCP).

runPercyScan writes

Run a Percy visual test scan (not available in Remote MCP).

fetchPercyChanges read-only

Get a summary of visual changes detected by Percy AI.

managePercyBuildApproval writes

Approve or reject a Percy build.

uploadProductRequirementFile writes

Upload a PRD/screenshot/PDF to get a file mapping ID (not available in Remote MCP).

createLCASteps writes

Generate Low Code Automation steps from a manual test case.

fetchSelfHealedSelectors read-only

Retrieve AI self-healed selectors to fix flaky tests.

prepareSelfHealingPlan read-only

Build a self-healing edit plan without modifying files.

fetchRCA read-only

Fetch AI Root Cause Analysis for failed tests (suggests fixes only).

getBuildId read-only

Get the BrowserStack build ID for a project and build name.

listBuildId read-only

Get the latest build ID across all users.

listTestIds read-only

List test IDs from a build, filtered by status.

Use cases

Manage test cases and runs using natural language.

Run manual tests on real browsers and devices.

Set up and run automated tests (Selenium, Appium, Playwright).

Debug test failures and get AI-powered root cause analysis.

Scan and fix web accessibility issues.

Integrate Percy visual testing and review changes.

Supported clients

Claude Desktop

GitHub Copilot

Cursor

Cline

Listed from the project's documentation, not tested by this site.

Overview

The BrowserStack MCP server is the official server from BrowserStack that brings the vast ecosystem of real device cloud and test platform into MCP clients such as Cursor, Claude, or GitHub Copilot. It enables managing test cases, executing manual or automated tests, debugging issues, and even fixing code using plain English commands. With 44 tools, it supports test management, automation test setup (Selenium, Appium, Playwright), observability troubleshooting, live device/browser testing, accessibility scanning, Percy visual testing, and BrowserStack AI agents.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →

NPM Sentinel MCP Server 79 · B

AI-powered NPM package analysis MCP server

★ 18 · Tools 19 Compare with this →

MartinLoop MCP Server 79 · B

An independent control layer that gives AI coding agents budgets, brakes, and receipts

★ 409 Compare with this →

GitHub MCP Server 76 · B

GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language

★ 33.5k · Tools 15 Compare with this →

Compare with other MCP servers

Source revision 58c829c6fc3c Data synced 2026-10-11 Read the FMRS scoring method