Cypress Cloud MCP for AI Agents. Audit E2E test failures via natural conversation.
Cypress Cloud MCP connects your AI agent directly to your end-to-end testing pipeline. Ask questions about test runs, pinpoint specific failures, track intermittent flaky tests, and generate detailed performance reports without logging into a dashboard. It gives you full conversational control over your entire E2E quality lifecycle.
Give Claude and any AI agent real-world access
Retrieves a list of organizational Cypress Cloud projects and their required unique IDs.
Provides status details for recent project runs, including pass/fail counts and commit information.
Pulls deep details on a specific failing spec file execution, providing error logs, video URLs, and screenshots.
Generates reports showing which tests fail randomly (flaky) and tracks the last time they failed across your code base.
Retrieves dedicated reports detailing test duration averages, p95 scores, and slow tests for CI/CD review.
Ask an AI about this
Waiting for input…
What AI agents can do with Cypress Cloud: 10 Tools for E2E Testing
These tools let you programmatically list projects, get run summaries, inspect specific failures, and generate comprehensive reports on test quality metrics.
Make your AI actually useful.
Add this MCP to Claude, Cursor, or Windsurf and your AI stops guessing. It gets real tools to look things up, take action, and handle the stuff you keep doing by hand.
Start using Cypress Cloud MCPList Projects
Lists all projects on Cypress Cloud; this helps you find the unique ID needed to run other commands.
Get Runs
Retrieves a summary of recent test runs, including their status and commit details.
Get Run
Gets comprehensive data for a single run, detailing total tests, passed/failed...
Get Instances
Lists every spec file execution within a given test run, showing statuses and...
Get Instance
Gets deep details on one specific spec instance, including error messages, videos...
Get Tests
Lists individual tests inside a single spec file execution, providing titles, status, and errors.
Report Runs
Generates an aggregated summary report of all test runs for a specific date range, useful for BI dashboards.
Report Tests
Creates a detailed enterprise report listing individual tests and their statuses...
Report Flaky
Identifies intermittently failing tests, returning names, flake rates, and last...
Report Slow
Generates a report on the slowest tests in your suite, detailing average duration...
Security and governance baked right in.
Pick your AI client below to get set up. Just create a Vinkius account, subscribe, and you're instantly up and running. We handle the entire backend infrastructure, delivering out-of-the-box support for HTTPS Streamable, SSE, and OAuth2—zero messy routing required.
Choose How to Get Started
Build a custom MCP for your own tools, or connect a ready-made integration from our catalog.
Build Your Own
Turn any API into an MCP. Import a spec, define Agent Skills, or deploy with MCPFusion.
- Import from OpenAPI, Swagger, or YAML specs
- Create Agent Skills with progressive disclosure
- Deploy to edge with MCPFusion framework
- Built in DLP, auth, and compliance on each call
- Real time usage dashboard and cost metering
- Publish to catalog or keep private
Make Your AI Do More
Start with Cypress Cloud, then connect any of our 5,200+ other servers whenever your AI needs more. One click, no limits.
- Use this MCP plus 5,200+ others, all in one place
- Add new capabilities to your AI anytime you want
- Connections are secured and governed automatically
- Track usage and costs across all your servers
- Works with Claude, ChatGPT, Cursor, and more
- New servers added to the catalog weekly
Independent Platform Disclaimer: Vinkius is an independent platform and is not affiliated with, endorsed by, sponsored by, verified by, or otherwise authorized by Cypress. All third-party trademarks, logos, and brand names are the property of their respective owners. Their use on this website is strictly for informational purposes to identify service compatibility and interoperability.
VINKIUS INFRASTRUCTURE
Cloud Hosted
Managed infra
V8 Isolated
Sandboxed per request
Zero-Trust Proxy
No stored credentials
DLP Enforced
Policy on each call
GDPR Compliant
EU data residency
Token Compression
~60% cost reduction
The pain of dashboard sprawl and manual correlation. Solved with Vinkius AI Gateway
Today, finding the root cause of a test failure is a multi-step scavenger hunt. You check the main Cypress Cloud dashboard for the run ID, then you copy that ID into another tab to view specs. If it failed, you click again to find the instance details, and finally, if you need proof, you have to export a CSV—all while juggling multiple tabs and copy/paste operations.
This process is slow, error-prone, and forces your agent to operate in silos. You spend more time navigating the UI than actually debugging code. The result is delayed fixes and missed performance issues because nobody has time to build a perfect reporting workflow.
Cypress Cloud MCP delivers deep audit data conversationally.
The Cypress Cloud MCP eliminates the need for manual correlation. You ask your agent, 'What were the slow tests in my latest run?' and it executes report_slow, pulling the precise average durations and p95 metrics without you ever leaving the chat window.
This changes everything. Instead of being a data analyst who builds reports, you're an auditor who gets instant answers. You talk to your quality assurance history as if it were just another API endpoint.
What your AI can actually do with this
Forget sifting through complex dashboards or exporting CSVs just to find out why a build failed. This MCP lets you talk to your testing results. You can ask your agent for an overview of recent runs, get the specific error messages and screenshots attached to a single failing test instance, or even request a report on flaky tests across your whole codebase.
The system pulls detailed information—like slow test metrics or aggregated run summaries—and presents it back to you in plain language.
Setting this up through Vinkius means you connect once, and your agent gets instant access to Cypress Cloud's full suite of audit tools. You get immediate visibility into performance gaps and quality trends, letting QA Engineers and DevOps teams stop guessing and start fixing the actual root cause.
019d7580-c770-70dc-9094-96c2ccda2db3 Here's how it actually works
The bottom line is you skip all the clicks. You ask the question, and the MCP pulls the necessary data from Cypress Cloud.
Subscribe to the Cypress Cloud MCP on Vinkius.
Enter your enterprise API Key and Project ID into your agent client.
Prompt your AI client with a natural language question, like 'Show me why the checkout process failed last week.'
Who is this actually for?
This connector is for QA Engineers who are tired of manual dashboard digging at 2 a.m. to find out why a critical test failed. It's for DevOps teams that need to prove CI reliability or developers who want immediate failure context without leaving their IDE.
Uses the MCP to inspect specific failing specs, retrieving error messages and video links to pinpoint why a test failed.
Runs reports on slow tests or generates aggregated run summaries to prove pipeline performance metrics for stakeholders.
Asks the agent to list individual test failures within a spec instance, getting immediate titles and error messages directly in their development workspace.
What Changes When You Connect
Debugging goes from clicking through dashboards to asking a question. Instead of manually cross-referencing run IDs and specs, you prompt your agent for the specific error messages or video links needed to diagnose a failure.
Stop treating flaky tests like random bugs. Use the report_flaky tool to automatically identify intermittent failures across the whole codebase and track exactly when they last occurred.
Get full performance visibility without exporting data. The report_slow tool pulls average, p95, and max durations instantly, giving you concrete metrics for optimizing your CI/CD pipeline reliability.
Audit compliance effortlessly. You can request a comprehensive run summary (report_runs) or individual test result data (report_tests), formatted right away for management review or BI dashboards.
Reduce context switching dramatically. Your agent lets you list projects and then drill down, moving from an overall status check (get_runs) to viewing granular error logs (get_instance) all in one chat thread.
See it in action
Debugging a critical failure immediately after merge.
A developer sees a test fail in the CI log. Instead of manually getting the run ID and then navigating to the instance, they ask their agent to 'Show me why instance X failed.' The MCP uses get_instance to pull the error logs, screenshots, and video links directly into the chat.
Reviewing quarterly stability for management.
An Engineering Manager needs a high-level view of quality. They prompt the agent to 'Generate an enterprise report on test results since Q1.' The MCP runs report_tests, providing aggregated data ready for a presentation slide.
Improving CI reliability by finding hidden issues.
A DevOps Engineer suspects tests are failing randomly. They ask the agent to 'Run a flaky test audit starting from last month.' The MCP uses report_flaky, giving them names and historical flake dates to prioritize fixes.
Investigating performance bottlenecks.
A team notices build times are creeping up. They ask the agent for 'a report on slow tests last week.' The MCP uses report_slow, providing specific metrics (avg/p95) and identifying which exact test methods need optimization.
The honest tradeoffs
What to watch out for, and the recommended way to handle each one.
Asking for general reports without scope.
Prompting: 'Give me the latest test report.' The agent will fail or give an unusable, vague summary because it doesn't know which time frame you mean. You get stuck in endless follow-up questions.
Always include necessary context. For run summaries, specify a date range and use the report_runs tool (e.g., 'Give me the enterprise run summary for 2024-06-01').
Diagnosing failures by guessing IDs.
A user finds an error log but doesn't know which project or instance ID to check next. They waste time trying different API endpoints, leading to frustration and incomplete data.
First run list_projects to get the required 6-character Project ID. Then use get_runs with that ID before attempting deeper dives.
Treating test status as a single boolean value.
Assuming 'passed' means everything is fine, when in fact there could be pending tests or slow warnings you need to see. You miss key data points.
Always check the run summary details using get_run, which provides total counts for passed, failed, and pending alongside the duration.
When It Fits, When It Doesn't
Use this MCP if your primary bottleneck is converting complex, structured testing data (run logs, performance metrics, error reports) into actionable insights without opening a dashboard. If you need to answer questions like 'What caused run 123 to fail?' or 'Which tests are flaky?', this is the tool for you. Don't use it if your goal is simply writing test code; that requires an IDE agent directly in your editor. Also, don't rely on this MCP for data outside of Cypress Cloud—it only talks to the API endpoints provided by Cypress. If you just need a list of all available projects, start with list_projects before querying any other report.
Questions you might have
How do I find the right Project ID using Cypress Cloud MCP? +
You start by calling list_projects. This tool retrieves a full list of all organizational projects, giving you the unique 6-character IDs needed to run any other report or query.
Can Cypress Cloud MCP tell me if a test is flaky? +
Yes. You use the report_flaky tool. It analyzes your history and generates reports identifying tests that fail intermittently, along with their last known failure dates.
What tools does Cypress Cloud MCP give for performance checking? +
It provides two main reporting tools: report_slow identifies the slowest methods by metrics like p95 duration, while get_run gives overall run summaries and counts that indicate general stability.
If a test fails, how do I use Cypress Cloud MCP to see the evidence? +
First, you might call get_runs for context. Then, use get_instances with the specific instance ID; this tool pulls the error messages, screenshots, and video URLs right into your chat.
Is Cypress Cloud MCP only for recent runs? +
No. You can generate historical reports using report_runs or report_tests by providing a specific start date (YYYY-MM-DD) to cover large time ranges for compliance audits.