The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →To connect an AI coding agent to a live Cypress session, run Cypress in open mode with Chrome, point Chrome DevTools MCP at the same remote debugging port that Cypress uses, and let the agent read the runner’s results together with the browser’s DOM, console and network data. Cypress also offers a separate terminal path, cypress tap, for agents that need Cypress’s own run status and failure details without a browser connection. This guide covers both, explains where Cypress Cloud MCP fits, and lists the security limits you should respect before you grant an agent access.
Choose the right approach before you configure anything
Three tools are often lumped together, but they answer different questions at different points in the workflow.
- Chrome DevTools MCP attached to Cypress gives the agent live access to the browser that Cypress is driving. Use it while you are developing or when you need to reproduce a failure locally.
cypress tapgives the agent a terminal interface to a runningcypress opensession. It returns Cypress’s view of a spec: status, the Command Log, the error and code frame, and the DOM at the point a command ran.- Cypress Cloud MCP works on recorded runs after CI has finished. It is for triage, not for driving a local browser.
Set up Chrome DevTools MCP to share Cypress’s browser
This setup depends on a matched port. Cypress’s documented approach is to pick a remote debugging port, configure Chrome DevTools MCP to connect to an existing Chrome instance on that port, and then start Cypress with the same value in CYPRESS_REMOTE_DEBUGGING_PORT. Cypress’s official example uses port 59210:
- Choose a free port number. The example in Cypress’s documentation is
59210. - Configure Chrome DevTools MCP to attach to an existing Chrome instance on that port. The exact configuration key depends on your MCP client, so follow the Chrome DevTools MCP documentation for your client rather than copying a flag from a blog post.
- From your project directory, start Cypress open mode with the same port:
CYPRESS_REMOTE_DEBUGGING_PORT=59210 cypress open --e2e --browser=chrome - Ask the agent to inspect the browser it can see. It should be looking at the Cypress-launched window and the Cypress runner, not a blank browser.
Cypress says this environment variable has been supported for many versions, so it is not a recent addition you need to check for. The critical requirement is that the two port values match exactly.
Troubleshooting a mismatched connection
- Symptom: the agent starts a fresh browser that has no knowledge of your Cypress spec. Cause: the MCP server did not reach the port Cypress opened. Fix: confirm the number is identical in both the MCP configuration and the
CYPRESS_REMOTE_DEBUGGING_PORTvalue, then restart Cypress and reconnect the agent. - Symptom: the agent sees a browser, but not the test you are running. Cause: Cypress was started in a different project directory or with a different spec. Fix: restart
cypress openfrom the project that contains the failing spec.
What the agent can see once connected
According to Cypress’s documentation, a connected agent can access:
- test pass or fail state and error messages;
- DOM state at the point of failure;
- console logs;
- network request data;
- Cypress command logs.
The practical loop Cypress describes is to ask the agent to inspect the latest run, combine those observations with the code and git history, decide whether the application or the test is wrong, apply a change, and then let Cypress rerun or reload as appropriate. Cypress’s illustrative example is a failing test involving deletion of a to-do item. That example shows the workflow, but it is not evidence of how often an agent fixes a failure correctly, and you should treat the outcome as something you verify yourself.
The terminal alternative: cypress tap
If you do not need the agent to drive a browser, cypress tap is a narrower option. Cypress documents it as an extension to its CLI that attaches to an open-mode Cypress session. Cypress states that an AI agent can run a spec and get back a pass or fail result, and that it can inspect the failing test’s Command Log, error and code frame, and the application’s DOM at the point a command ran.
The documented flow is:
- In your project, run
cypress open. - Select a testing type and a Chromium-family browser.
- From another terminal in the same project, issue
cypress tapcommands.
Cypress recommends the --json output for agents and scripts. It is included with the Cypress App and, according to Cypress, requires no Cloud account or paid subscription.
Limits of cypress tap
- It requires Cypress v15.21.0 or later.
- It attaches only to
cypress open, not to headlesscypress run. - It supports Chromium browsers: Chrome, Chromium, Edge and Electron.
- It is marked beta, and its commands and output may change in a future release.
Local live debugging versus Cypress Cloud MCP
Chrome DevTools MCP and cypress tap work on your machine during development. Cypress Cloud MCP works after a run has been recorded in Cypress Cloud. An agent connected to Cloud can query run status, flaky tests, failure details and Test Replay links. Use it when a CI failure needs investigating and you do not have a local reproduction yet.
Cypress states that Cloud MCP reached general availability on May 20, 2026, and that it is included on every Cypress Cloud plan at no additional cost. Access requires an organization admin to enable the integration, and each user must authenticate. Cypress recommends OAuth, and documents personal access tokens as an alternative. Plan inclusions and authentication options are service terms that can change, so confirm them in your Cloud organization settings before you rely on them.
Compare the options
| Option | Best fit | Setup and access | Important limits |
|---|---|---|---|
| Chrome DevTools MCP attached to Cypress | Live browser state, DOM, console and network inspection alongside the Cypress runner | Match the Chrome remote debugging port in the MCP configuration and in CYPRESS_REMOTE_DEBUGGING_PORT |
Requires an open-mode Chrome session and a matched port; grants privileged browser access (Cypress documentation; Chrome for Developers) |
cypress tap |
Agent-readable Cypress status, Command Log, error details and failure-time DOM from the terminal | Start cypress open, then run CLI commands from another terminal in the project |
Cypress v15.21.0 or later; Chromium browsers only; open mode only; beta (Cypress documentation) |
| Cypress Cloud MCP | Triage of recorded CI runs, including run, test and replay information | Enabled by an organization admin; each user authenticates (OAuth recommended, personal access tokens as an alternative) | Works on recorded Cloud runs rather than a local live browser; generally available from May 20, 2026 per Cypress; plan and auth terms may change |
Security before you connect an agent
Chrome for Developers warns about this kind of connection directly: “Because your agent will be able to view and interact with the pages it accesses, it can effectively act on your behalf if you connect it to a browser with an active, authenticated session.” In practice, the agent can read, inspect, debug and modify browser and DevTools data.
- Use a test account, not a personal or production account.
- Do not expose sensitive personal data or production data to the test browser.
- Remember that Cypress launches its own browser profile, separate from your everyday browser. Cookies and extensions from your normal profile do not carry over automatically, so you should sign in to the test environment inside the Cypress-controlled session when a test requires it.
Cypress describes cypress open as headed and interactive, and cypress run as headless by default. A live-browser connection is therefore for interactive debugging, not for unattended CI runs.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsWhere this leaves you
Use the matched-port setup when you want an agent to inspect the browser while a spec is failing on your machine. Use cypress tap when you want a terminal-based, Chromium-only view of Cypress’s own results. Use Cypress Cloud MCP when the failure only exists in a recorded CI run. Verify version requirements, the beta status of cypress tap, and Cloud plan and authentication details against Cypress’s current documentation before you roll this out to a team.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
