browserstack/mcp-server

Fetch build details and quality gate results. Includes hashed_id and session_type for listSessions.

Retrieves and summarizes all visual changes detected by Percy AI between the latest and previous builds, helping quickly review what has changed in your project.

Description is built at runtime.

Fetch BrowserStack self-healed selectors plus the test source code for the run. Provide exactly one of `sessionId` (single Automate / App-Automate session) or `buildUuid` (full self-healing report for a build). Pass the returned locator pairs to `prepareSelfHealingPlan` to plan edits. Requires a session id. listSessions lists them for a build; if you do not have the build either, getBuildId or listBuildId resolves one from a project and build name.

Retrieve an existing authentication configuration by ID. Id: from createAccessibilityAuthConfig, or askBrowserStackAI with product "a11y" to find one.

Get a build's server-computed failure-theme clusters (buildThemes + buildWorkflows), the preferred grouping source for representative/sibling fan-out — call this once per build, not per test. Triggers server-side computation if nothing has ever run for this build, and polls in-call up to its own budget for the result to finish; never blocks indefinitely. ready:false means either the computation is still running past the poll budget or the trigger itself failed (status:'trigger-unavailable') — either way, fall back to client-side clustering rather than waiting longer or retrying this call.

Get the BrowserStack build ID for a given project and build name, scoped to the current user's builds.

Fetch logs, or the session video URL, for an Automate/App Automate session. Requires a session id. listSessions lists them for a build; if you do not have the build either, getBuildId or listBuildId resolves one from a project and build name.

Use this tool to debug failures in the last run of the test suite on BrowserStack. Use only when browserstack.yml file is present in the project root.

Description is built at runtime.

Description is built at runtime.

Read a previously submitted RCA turn's status once, given the testRunId and the turnId a PENDING tfaRcaTurn response returned — this never resubmits or duplicates the turn. Returns PENDING again if the TFA agent is still working, or the same RESOLVED/NEEDS_INFO contract as tfaRcaTurn once it finishes. On NEEDS_INFO, gather the requested evidence and continue via tfaRcaTurn with the same threadId; RESOLVED is terminal for this test — do not call this tool or tfaRcaTurn again for it.

Call a capability whose contract you have from describeCapability. Pass `name` exactly as given — that is the handle. Only when a result carries no `name` (some products do not publish them yet) pass `method` and `path` instead, exactly as returned. Arguments go in path_params / query / body under the spec's own names. One call makes exactly one request and returns the product's own response untouched; when `completed` is false there is another page, which you fetch by sending the capability's own page parameter. If the mode is 'write' you MUST ask the user first, then resend with user_permiss…

Get the latest build ID for a project and build name, across all users (no user filter).

List folders in a BrowserStack Test Management project, returning each folder's id and name (plus case counts and sub-folder counts). Pass parent_id to list sub-folders under a specific folder instead of top-level folders. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.

List the BrowserStack products this surface can reach: what each one does, the entities it models, and a line saying what each entity is. START HERE — searchCapability needs a product, and this is what tells you which one. If two products could both fit the task, ask the user rather than choosing for them. ALWAYS pass the user's request, in their own words, as `query`: the answer then says outright whether the products clash over the words in it. WHEN IT REPORTS A CLASH — two products claiming the same word for different things — ASK THE USER WHICH PRODUCT THEY MEAN AND WAIT FOR THEIR ANSWER B…

List sessions for a hashed Automate/App Automate build: session IDs, status, OS, browser/device, dashboard URL.

Description is built at runtime.

List test cases in a project, optionally scoped to a specific folder. Omit folder_id to list all test cases in the project; provide folder_id (discoverable via listFolders) to list only that folder's cases. Supports filters: case_type, priority, pagination. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.

List test-case templates with their numeric template_id. Use the id with createTestCase to apply a custom template (the 'template' slug only selects system templates).

Lists all test files for a given set of directories.

Description is built at runtime.

List test plans in a BrowserStack Test Management project. Returns each plan's identifier (TP-*), name, status, description, dates, and active/closed test-run counts. Supports pagination. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return. Creating a test plan or sub-plan, and linking or unlinking test runs on one, are not available as tools here: use the capability functions, which reach Test Management's own API for these operation…

List test runs in a project with optional filters (date ranges, assignee, state, etc.) Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.

List the test runs belonging to one failure theme or workflow from a prior getBuildFailureThemes call, filtered by themeId or workflowId. Results are paginated — pass the previous response's nextCursor back as cursor and keep calling until no nextCursor is returned; do not assume a single page covers all members. This is the representative/sibling grouping source: every test in a theme's list becomes either the cluster's representative or a pre-seeded sibling confirm.

Approve or reject a Percy build

Integrate Percy visual testing into new projects and demonstrate visual change detection through a step-by-step simulation. This tool handles initial Percy integration, setup, and creates a working demonstration for new users. Primary tool for prompts like: "Integrate percy for this project", "Integrate percy in this project {project_name}"

Build a self-healing edit plan that bundles locator pairs with test source code for the calling LLM to apply with its own editing tools. This tool does NOT modify files — it returns structured context so the LLM can edit only the relevant call sites (avoids blind find/replace across shared ids/classes). `sessions` accepts the canonical shape `[{sessionId, locators: [{original, healed, thought?}]}]`, a single session object, an envelope `{sessions: [...]}`, the raw report `{healing_logs: [...]}` (with `healed_selectors` aliasing `locators`), and snake_case keys (`session_id`, `original_locator`…

Use this tool when user wants to manually check their app on a particular mobile device using BrowserStack's cloud infrastructure. Can be used to debug crashes, slow performance, etc.

Execute pre-built native mobile test suites (Espresso for Android, XCUITest for iOS) by direct upload to BrowserStack. ONLY for compiled .apk/.ipa test files. This is NOT for SDK integration or Appium tests. For Appium-based testing with SDK setup, use 'setupBrowserStackAppAutomateTests' instead.

Launch a BrowserStack Live session (desktop or mobile).

Run a Percy visual test scan. Example prompts : Run this Percy build/scan. Never run percy scan/build without this tool

Find endpoints this surface can call, by plain language, within ONE product. CALL listProducts FIRST, passing the user's request as its `query`: it names the products, says whether they clash over the words in that request, and gives you the question to ask when they do. You need it anyway — `product` here takes a name exactly as listProducts spells it. Where two products use the same word for different things, ask the USER and wait for an answer rather than picking. Search matches the product's OWN words, not synonyms. When your words are not the product's, the response says so: `weak_match: …

Set up BrowserStack App Automate SDK integration for Appium-based mobile app testing. ONLY for Appium based framework . This tool configures SDK for various languages with appium. For pre-built Espresso or XCUITest test suites, use 'runAppTestsOnBrowserStack' instead.

Set up and run automated web-based tests on BrowserStack using the BrowserStack SDK. Use this tool for functional or integration test setup on BrowserStack only. For any visual testing or Percy integration, use the dedicated Percy setup tool. Example prompts: run this test on browserstack; set up this project for browserstack.

Scan a web page URL for accessibility issues and get a CSV report. For a mobile app, run it on App Automate with accessibility on (setupBrowserStackAppAutomateTests), then fetchAccessibilityIssues.

Report a blocker, tool error, missing capability, or user feedback to BrowserStack. Never include PII or credentials.

Use this tool to take a screenshot of an app running on a BrowserStack device. This is useful for visual testing and debugging.

Submit one collaborative RCA turn for a test run to the TFA agent, then poll in-call for a result. Omit threadId on the first turn for a test run — this starts a new investigation thread; every following turn on that same test passes back the threadId from the previous response, since a test run has only one active thread at a time. Returns status RESOLVED (terminal, includes root_cause/related_prs), NEEDS_INFO (fulfill the asks and submit the next turn with the same threadId), or PENDING (still working — poll it with getTfaTurnResult using the returned turnId; calling tfaRcaTurn again for the…

Trigger (or read, if one already exists) a build's Release Readiness report, returning a trimmed verdict glimpse and a Test Observability UI link — never the raw report body. Without force, a completed report is read as-is at no extra analysis cost, so it's safe to call this repeatedly just to check status. Pass force:true only to force a fresh re-analysis of an already-completed report (e.g. after new evidence changed the picture) — do NOT set force on every call, since that discards the cached report and always pays the full re-analysis cost.

Update an existing test case in BrowserStack Test Management. Any subset of the following fields may be changed: name, description, preconditions, test_case_steps, owner, priority, case_type, automation_status, status, tags, issues, custom_fields. Only the supplied fields are modified. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.

Update a test run's metadata and/or add test cases to it. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.

Upload files (e.g., PDRs, PDFs) to BrowserStack Test Management and retrieve a file mapping ID. This is utilized for generating test cases from files and is part of the Test Case Generator AI Agent in BrowserStack. Requires a project identifier (PR-*). No tool here lists projects, so if you do not have one, use the capability functions to list the Test Management projects, then retry this tool with the identifier they return.