Running Cypress?

Cypress tests flaky and nobody trusts them? We bring it back.

A Cypress suite that spent a year accumulating cy.wait calls and defaultCommandTimeout bumps stops being a test suite and becomes a wall of ignored red. We read the retries, remove the waits that are hiding real bugs, and leave you with a shorter suite you can genuinely rely on at the end of a release.

  • Since 2015Fixing technology
  • FreeEstimates
  • US teamReal engineers

cypress.log

still flaky

What it is

Cypress at a glance

Cypress is a JavaScript end-to-end testing framework for web applications, released by Brian Mann and Cypress.io in 2017. It runs tests inside the browser itself rather than driving one remotely, which gives it direct access to the application's state and a runner window where tests play back step by step. Along with the open-source runner, Cypress Cloud provides a dashboard, parallelisation and flake detection as a paid service.

Cypress 10 restructured how tests are configured, Cypress 12 removed the experimentalSessionAndOrigin flag and made cy.session and cross-origin support production features, and recent versions added better component testing for React, Vue, Angular and Svelte. The runner is tightly tied to Electron and Chromium, and Firefox and WebKit coverage are less polished than in Playwright.

Type
End-to-end and component testing framework
First released
2017
License
MIT for the open-source runner, paid service for Cypress Cloud
Maintained by
Cypress.io, Inc.
Typically used for
End-to-end web testing, component testing for React and Vue, authenticated user flows, API plus UI mixed suites, CI smoke tests

The honest version

Cypress pros and cons

No technology is all upside. This is where Cypress earns its place and where it tends to cause trouble.

Where Cypress is strong

  • The runner window is genuinely usefulWatching a test replay step by step with DOM snapshots at each command is still the clearest debugging experience in the E2E space for anyone new to the tool.
  • Writing a first test is quickThe API reads top to bottom, matchers chain cleanly, and a working happy-path test is a page of code rather than a project.
  • Great for component tests on React and VueComponent mode gives you a real browser render with the same assertions and spies, which catches integration bugs unit tests miss.
  • cy.intercept is a strong mocking APIIntercepting network requests, modifying responses and asserting on calls happens in one place, which makes contract-style tests tractable.
  • Cypress Cloud handles orchestrationParallelisation, retries, flake detection and screenshot storage are a paid service away, and the integration with GitHub Actions is well documented.
  • A huge body of existing tests exists in the wildHiring for Cypress experience is still easier than for most alternatives, and example tests for common frameworks are everywhere.

Where Cypress bites

  • Cross-browser coverage is weakFirefox support lags and WebKit is not supported at all in a meaningful way, so Safari regressions still have to be caught somewhere else.
  • The architecture has limitsRunning in the browser means tabs, multiple origins, native dialogs and some iframe patterns are either awkward or impossible without workarounds.
  • cy.wait culture ruins suitesTeams reach for cy.wait and defaultCommandTimeout bumps to paper over flakes, and the suite accretes dead weight that hides real problems.
  • Cypress Cloud costs scale fastParallel machines, test runs and retention tiers add up, especially when a flaky suite is retrying every test three times. The bill is quiet until it is not.
  • Slow suites stay slowEven with parallelisation, a mature Cypress suite with hundreds of tests often runs for twenty minutes or more, and gating every merge on it costs real development time.
  • Momentum has moved to PlaywrightMany new projects start on Playwright instead, and the volume of new Cypress plugins and articles is lower than it was. That matters for finding answers to unusual problems.

Sound familiar?

Common Cypress problems we fix

The Cypress problems that bring people to us, and what you have once each one is fixed.

cy.wait timed out waiting 30000ms

The suite is full of cy.wait calls

A previous team reached for cy.wait every time something flaked, and now every test blocks on fixed timeouts that hide the real behaviour.

Waits replaced with assertions

Flake detected (3 retries)

Retries are masking real bugs

Cypress Cloud is retrying failing tests three times by default, which hides regressions behind a green badge until a customer surfaces them in production.

Retries tuned, real failures visible

cy.intercept: no request matching

Mocks silently stop matching

A URL pattern no longer matches after an API refactor, so cy.wait on the intercept hangs forever and the test reports a timeout rather than a bad mock.

Intercepts pinned, mocks honest

Cypress Cloud: $4,240/month

Dashboard spend got away from the team

A combination of extra parallel machines, retention tiers and flaky test retries has quietly doubled the Cypress Cloud invoice over a couple of quarters.

Spend aligned with real suite needs

cy.session not migrated

Session handling is stuck on deprecated APIs

experimentalSessionAndOrigin was removed in Cypress 12, and the suite is still using the old shape. Login steps are being repeated per test at a huge time cost.

cy.session in place, login once per spec

Chromium Renderer process just crashed

Browser crashes mid-suite

Memory leaks from long specs, large payloads being captured in snapshots, or extensions loaded into the test browser are pushing it over the edge during CI runs.

Memory contained, browser stable

no ownership, 2,000 tests

The suite has become unmaintainable

A sprawling tests directory, a lot of green noise, a smaller amount of red that everybody ignores, and nobody whose job it is to look at the Dashboard every morning.

Suite trimmed, ownership defined

cross-origin iframe assertion fails

Cross-origin flows are fragile

A test that touches an OAuth provider or an embedded Stripe iframe falls apart because cross-origin support was never configured correctly for the suite.

Origins configured, flows end to end

What we do

Cypress help, from flake rescue to replacement

Call us for one unstable suite, or let us take the whole testing pipeline in hand.

Rescue
A suite you cannot trust, a CI run that retries every test, or a Dashboard full of red. We stabilise what matters and quietly delete the rest.
Flake triage
Classifying which failures are the app being honest, which are test code written against a race, and which are infrastructure issues in CI.
Cypress 9 to 12+ upgrades
Config file conversion, cy.session migration, component testing updates and runner configuration verified against the current major.
Move to Playwright
Where the architecture is holding you back, we port the suite to Playwright, keeping coverage and dropping the parts that no longer earn their keep.
Cypress Cloud spend review
A look at parallelisation, retries and retention, with recommendations that usually trim the invoice without losing CI throughput.
Managed Tech
Day-to-day care of the suite, a pair of eyes for every new flake, and minor Cypress upgrades handled without interrupting the team.

How it works

From first message to fixed

  1. Look Free estimate

    Tell us what's going wrong. We look at it, then tell you plainly what's wrong and what it will cost.

  2. Fix

    We fix the cause, not the symptom. If something is down, getting it back up comes first.

  3. Harden

    Backups, monitoring, updates and access, so the same problem can't quietly come back.

  4. Hand back or stay on

    You get a plain-language write-up. Or we stay on for hosting, upkeep and new work.

Often found together

Running Cypress alongside something else?

We work on the whole stack, so the problem can be wherever it actually is.

Questions

Cypress questions

Something else? Call+1 737 727 7327

Can you stabilise our flaky Cypress suite?

Yes, and this is the most common job we take on. We read the retry history from Cypress Cloud or CI, classify each flake as a real bug, a test-code race or an infrastructure issue, and apply the right fix for each. Most suites come back to under a one percent flake rate within a week or two.

Should we switch from Cypress to Playwright?

Sometimes. If you need WebKit coverage, better cross-origin support, or a faster parallel runner, Playwright is a clear win. If your suite is mostly Chromium flows and the team knows Cypress well, we often get more value from stabilising what you have than from a full migration.

Why is our Cypress Cloud bill so high?

The common causes are more parallel machines than the suite actually needs, retries left on for every test, and generous retention of recordings. We review the Cloud usage, show you where the money is going, and typically reduce the invoice without slowing CI down meaningfully.

Can you add Cypress to our CI pipeline?

Yes. We wire it into GitHub Actions, GitLab CI or CircleCI with the right Docker image, cache configuration, parallelisation, artifact uploads and failure screenshots. If you want Cypress Cloud it integrates on top; if you prefer to stay on the open-source runner, we keep the setup pure.

Our old Cypress suite uses cy.wait everywhere. Can you fix it?

Yes. We replace cy.wait calls with assertions on application state or network interceptions that resolve when the work is actually done. The suite usually gets faster and more reliable at the same time, because the waits were covering for something specific we now handle directly.

How much does Cypress help cost?

The quote itself is at no charge. Share a recent CI run and the config file, and we come back with a number before any changes are proposed.

Free estimate

Cypress flaking on every release?

Tell us what you're running and what's going wrong. A person reads it and comes back with an estimate.

  • No charge to find out what it will take
  • If it's urgent, say so

Replies come from a person. By sending, you authorize us to email you back. No marketing.