BrowserStack MCP Server: Setup for Claude, Cursor and Other Apps

• Lists BrowserStack products and what they cover

  Names the BrowserStack products and their entities. It is the first step before searching for a capability.

  Try: “What can I do with BrowserStack?”

  Read · listProducts

• Finds an action in one product by plain description

  Looks up capabilities inside a single BrowserStack product from a plain-language request.

  Try: “Find how to start a load test.”

  Read · searchCapability

• Shows what an action needs before you run it

  Returns the full contract of a capability, such as its parameters, before it is run.

  Try: “What details does that capability need?”

  Read · describeCapability

• Explains one BrowserStack item and its identifiers

  Describes an entity, including the identifiers used to refer to it.

  Try: “Explain what a test plan ID looks like.”

  Read · describeEntity

• Runs a BrowserStack action found by search

  Runs a capability after it has been described. It can create or change items, for example in Test Management. The source says deletes are refused and that writes need your approval first.

  Try: “Stop the running load test.”

  Write · invokeCapability

• Opens a website in a real browser for you to use

  Starts a manual Live browser session on a browser, version and OS you choose, including local sites. You operate it by hand, the assistant cannot drive it.

  Try: “Open localhost:3001 on the latest Edge.”

  Runs code · runBrowserLiveSession

• Opens your app on a real phone for you to use

  Starts a manual App Live session on a real device. You operate it by hand, the assistant cannot drive it.

  Try: “Open my app on an iPhone 15 Pro Max.”

  Runs code · runAppLiveSession

• Takes a screenshot of an app on a BrowserStack device

  Captures a screenshot of the app running on a cloud device.

  Try: “Screenshot the app on that device.”

  Read · takeAppScreenshot

• Sets up your web tests to run on BrowserStack

  Gives the steps and config to run an existing web test suite on BrowserStack Automate.

  Try: “Set up my Playwright tests to run on BrowserStack.”

  Write · setupBrowserStackAutomateTests

• Sets up your mobile app tests to run on BrowserStack

  Gives the steps and config to run a mobile app test suite on BrowserStack App Automate.

  Try: “Set up my Appium tests for BrowserStack.”

  Write · setupBrowserStackAppAutomateTests

• Runs your mobile app tests on BrowserStack devices

  Starts your app tests on real devices in App Automate.

  Try: “Run my Android tests on a Pixel 8.”

  Runs code · runAppTestsOnBrowserStack

• Finds the likely cause of a failed test

  Returns an AI root-cause analysis for failed Automate or App Automate tests. It only suggests fixes and never changes your code on its own.

  Try: “Why did the checkout test fail?”

  Read · fetchRCA

• Gets logs or the video of a test session

  Fetches the logs or the video link for an Automate or App Automate session.

  Try: “Show the logs for that failed session.”

  Read · getFailureLogs

• Gets screenshots from a test session

  Returns screenshots captured during an Automate or App Automate session.

  Try: “Show screenshots from session abc123.”

  Read · fetchAutomationScreenshots

• Finds the latest build for a project

  Returns the latest build ID for a project and build name.

  Try: “Get the latest build for project Shop.”

  Read · listBuildId

• Finds a build ID by project and name

  Returns the BrowserStack build ID for a given project and build name.

  Try: “What is the ID of the nightly build?”

  Read · getBuildId

• Lists the tests in a build

  Lists every test of a build with its status and session ID, with an optional status filter.

  Try: “List the failed tests in that build.”

  Read · listTestIds

• Lists the sessions of a build

  Returns the sessions for an Automate or App Automate build, with their IDs.

  Try: “List sessions in that build.”

  Read · listSessions

• Shows build details and quality gate results

  Returns build details and the results of quality gates for a build.

  Try: “Did the last build pass its quality gates?”

  Read · fetchBuildInsights

• Finds selectors BrowserStack repaired itself

  Returns the self-healed selectors and the test source for one session or build.

  Try: “Which selectors were self-healed in this build?”

  Read · fetchSelfHealedSelectors

• Plans edits to adopt the healed selectors

  Builds an edit plan from healed selector pairs and the test source. It does not change any files.

  Try: “Plan the selector fixes for my tests.”

  Read · prepareSelfHealingPlan

• Scans a page or app for accessibility problems

  Scans a web page URL, or a mobile app, for accessibility issues and returns a CSV report.

  Try: “Scan www.example.com for accessibility issues.”

  Network · startAccessibilityScan

• Fetches issues from an accessibility scan

  Returns accessibility issues from one scan or report source.

  Try: “Show the issues from that scan.”

  Read · fetchAccessibilityIssues

• Answers accessibility questions

  Answers questions about accessibility, such as WCAG rules.

  Try: “What does WCAG say about contrast?”

  Read · accessibilityExpert

• Saves a login for scanning private pages

  Creates an authentication setup for accessibility scans, using a form login or basic authentication.

  Try: “Create a login config for my staging site.”

  Write · createAccessibilityAuthConfig

• Shows a saved scan login

  Returns an accessibility auth configuration by its ID.

  Try: “Show scan login config 42.”

  Read · getAccessibilityAuthConfig

• Creates a Test Management project or folder

  Adds a project and/or folder in BrowserStack Test Management.

  Try: “Create a project My Demo Project with Login and Checkout folders.”

  Write · createProjectOrFolder

• Lists folders in a project

  Lists folders in a Test Management project.

  Try: “List the folders in My Demo Project.”

  Read · listFolders

• Lists test cases in a project

  Lists test cases in a project, with filters.

  Try: “List high priority Login test cases.”

  Read · listTestCases

• Adds a test case

  Creates a test case in Test Management.

  Try: “Add an invalid login test case.”

  Write · createTestCase

• Edits a test case

  Changes only the fields you supply on an existing test case.

  Try: “Set the invalid login test to high priority.”

  Write · updateTestCase

• Lists test case templates

  Lists templates with their numeric IDs, to apply a custom template when creating a test case.

  Try: “Which test case templates exist?”

  Read · listTestCaseTemplates

• Writes test cases from a document

  Generates test cases from a file with Test Management's Test Case Generator AI agent.

  Try: “Generate test cases from this requirements file.”

  Write · createTestCasesFromFile

• Uploads a requirements file

  Uploads a product requirements file for test case generation.

  Try: “Upload the PRD for test case generation.”

  Write · uploadProductRequirementFile

• Writes Low Code Automation steps for a test case

  Generates Low Code Automation steps for a test case.

  Try: “Create automation steps for the login test case.”

  Write · createLCASteps

• Creates a test run

  Adds a test run in Test Management.

  Try: “Create a test run for the Login tests.”

  Write · createTestRun

• Lists test runs

  Lists test runs, with filters such as dates.

  Try: “List test runs from this week.”

  Read · listTestRuns

• Edits a test run or adds test cases to it

  Updates a test run or adds test cases to it.

  Try: “Add the checkout tests to run 12.”

  Write · updateTestRun

• Records a test result

  Adds a test result to a test run.

  Try: “Mark the login test as passed in run 12.”

  Write · addTestResult

• Lists test plans

  Lists test plans with their status.

  Try: “List our test plans.”

  Read · listTestPlans

• Shows a test plan

  Returns the metadata of a test plan.

  Try: “Show test plan TP-3.”

  Read · getTestPlan

• Lists a test plan's sub-plans

  Lists sub-plans under a test plan, in pages.

  Try: “List sub-plans of TP-3.”

  Read · listSubTestPlans

• Shows a sub-plan and its runs

  Returns a sub-plan with its metadata and linked runs.

  Try: “Show sub-plan STP-5.”

  Read · getSubTestPlan

• Runs a Percy visual test

  Starts a Percy visual scan of your project.

  Try: “Run a Percy scan.”

  Runs code · runPercyScan

• Adds Percy snapshot calls to your tests

  Gives instructions to add Percy snapshot commands to your tests.

  Try: “Add Percy snapshots to my tests.”

  Write · addPercySnapshotCommands

• Sets up or widens Percy visual testing

  Gives setup steps to add Percy visual testing to a project.

  Try: “Set up Percy for my project.”

  Write · expandPercyVisualTesting

• Simulates a visual change to try Percy

  Helps simulate a visual change so you can see Percy detect it.

  Try: “Simulate a UI change to test Percy.”

  Write · percyVisualTestIntegrationAgent

• Summarizes visual changes Percy found

  Summarizes the visual changes Percy AI found between the latest and previous builds.

  Try: “What changed visually in the latest build?”

  Read · fetchPercyChanges

• Approves or rejects a Percy build

  Marks a Percy build as approved or rejected.

  Try: “Approve the latest Percy build.”

  Write · managePercyBuildApproval

• Lists your local test files

  Lists test files in your project to run on BrowserStack.

  Try: “Which test files can I run?”

  Files · listTestFiles

• Hands a task to BrowserStack's own assistant

  Sends a plain-language task to BrowserStack's own agent and passes its permission requests back to you to approve. Its description says deletes are refused.

  Try: “Ask BrowserStack to set up a test run for me.”

  Write · askBrowserStackAI

• Used by the assistant behind the scenes

  Sends a feedback report to BrowserStack

  Lets the assistant report a blocker, tool error or missing capability to BrowserStack, in up to 2000 characters. The description says never to include personal data or credentials.

  Try: “Tell BrowserStack the Percy tool failed with a timeout.”

  Network · submitFeedback

Watch out for

• Deletes are refused in the source

  The capability tools write to your BrowserStack data but refuse destructive endpoints. The source comment says the invoke tool "can never delete", and writes need your approval first. Archive, close and merge count as writes. npmjs.com

• It changes your BrowserStack data

  Tools create and edit test cases, runs and plans, and approve or reject Percy builds. Use an account scoped to the project you are working on. github.com

• It can edit your test setup

  The setup tools produce config and code changes for your test projects. Review the diff before accepting. github.com

• Fix suggestions are not applied for you

  The README says root-cause results are proposals only, and your assistant must ask before editing code. github.com

Security advisories