Skip to main content
Free plan, no cardStart free

REGRegression testing · updated 2026-10-27

Regression testing tool with the evidence of every rerun

Record a flow once and rerun it as a regression test on demand or on a schedule, with smart selection, visual baselines, API monitors and every rescue shown.

For: QA leads and SDETs who own a regression suite that has to keep passing while the interface changes, and teams that want UI, visual and API regression checks to read the same environments and credentials.

acme-shop / checkout-flow.recIllustrative run
  1. REGRegressionPassed14 / 14 steps · 38.2 s
  2. LOADLoad · 200 virtual usersIn budgetP50 180 · P95 412 · P99 690 ms
  3. SECSecurity · OWASP Top 10 (2025)1 confirmed1 confirmed · 2 observed
  4. A11YAccessibility · WCAG 2.2 A and AA3 issues9 pages · VPAT draft ready
1 selector healed#pay-now-btn[data-test="pay"]Shown in the run

Capabilities

Your recordings are the suite

A flow recorded once in a real browser replays as a regression test straight away. Group flows in folders and run them together on a schedule, or by hand after a release.

Read more: Your recordings are the suite

Run what a change can reach

Smart test selection maps a change or a requirement to the flows that can reach it and runs those within a time budget, with a written rationale per pick and a confidence score that says when to run everything.

Read more: Run what a change can reach

Visual regression beside functional

A baseline screenshot per page and viewport, a pixel diff against the threshold you set and element drift on the DOM catch a collapsed layout or a hidden button that every functional assertion would pass.

Read more: Visual regression beside functional

API regression on a schedule

API collections run in the Collection Runner on demand or on a schedule as a monitor, with assertions on status, headers, body and schema and microsecond timing, so a slower endpoint is a recorded change.

Read more: API regression on a schedule

Rescues you can audit

When a locator breaks, the other locators stored at record time try to rescue the step. The run shows the original, the replacement and the strategy that found it, so you can check that a rescue did not hide a real change.

Read more: Rescues you can audit

Evidence on every run

Each run records pass or fail per step with its screenshot and assertion result, and the recording behind it keeps the seven structured files of its session, from element details to the replayable steps.

Read more: Evidence on every run

How it works

  1. Record the journeys that must not breakPoint the recorder at a URL you are allowed to test, click through each journey once and add assertions for what each step must prove.
  2. Group them and set a timetablePut the flows in a folder and run it on a schedule, or run it by hand after a release.
  3. Add baselines where the layout mattersCapture a visual baseline per page and viewport and set the pixel threshold a change may cross.
  4. Let selection trim the runWith smart test selection, a change runs only the flows it can reach, each pick with its written reason.
  5. Read what changedA failed step shows its screenshot, network calls and assertion; a rescued step shows the replacement locator beside the original.

What you receive, and what we do not do

Evidence ledger
Per run
Pass or fail per step, duration, browser, and the same flow exportable as Playwright, Selenium or Robot Framework code.
Per failed step
The element targeted, the assertion result, a screenshot and the network calls the step caused.
Per rescue
The locator that failed, the one that matched, the strategy that found it and its confidence.
Per selection
The flows picked for a change, the written rationale for each and a confidence score.
Per visual check
The difference as a percentage against your threshold, and a highlighted diff image.

Limits we state

  • There is no CI trigger, CI token or quality gate yet: run the suite in the platform on demand or on a schedule, or export the flows as code and run them in your own pipeline.
  • Scheduled folder runs use Chromium; runs across several browsers go through the execution grid.
  • Locator rescues are accepted during the replay (folder runs require a confidence threshold first), and there is no customer review queue yet: check the steps a run marks healed.
  • Running API collections from CI is not available yet; they run in the platform on demand or as scheduled monitors.
  • We publish no maintenance-saving or flakiness percentage. The run history shows the numbers for your own suite.

Plans: Recording and replay are on every plan, including Free (10 recordings a month). Folders, smart test selection and visual regression start at Solo Pro, the execution grid at Team. API collections run on every plan within its API test-run quota. See the metering table.

Questions buyers ask

What is a regression testing tool?

Software that reruns the tests which once proved a feature worked, after each change, and reports what broke. Recording-based tools build those tests from a flow you click through once; code-based frameworks such as Selenium and Playwright need each test written by hand.

Can I run the regression suite in my CI pipeline?

Not from the platform yet: there is no CI trigger, CI token or quality gate. Export the flows as Playwright, Selenium or Robot Framework code and run them in your pipeline, or run them in the platform on demand or on a schedule.

How does it decide which tests to run after a change?

Smart test selection, on Solo Pro and above, maps the change or the requirement to the flows that can reach it and runs those within a time budget. Each pick carries a written rationale and the selection a confidence score; when the confidence is low, run the whole suite.

Does it do visual regression testing?

Yes, from Solo Pro: a baseline screenshot per page and viewport, a pixel diff against the threshold you set, and element drift on the DOM, so a layout change that passes every functional assertion is still caught.

What happens to the suite when the interface changes?

A step whose locator breaks tries the other locators stored at record time. If one matches, the run continues and shows the rescue beside the original; if none matches, the step fails with its screenshot, and that journey is recorded again.

What should a regression suite contain?

The journeys that earn money or trust: signing up, signing in, paying, and the main thing the product does. Then every flow a past incident touched, because a defect that happened once is the one most likely to happen again. Keep the suite as small as it can be while still covering each change; a flow that never fails and never changes belongs in the scheduled run, not in the one that gates a release.

How often should regression tests run?

The flows a change can reach, on every change; the whole suite on a schedule, nightly or before each release, so a defect is found the day it is made rather than the week it ships; and the API checks that guard an integration on a fixed interval, independent of deploys. In Dokimore the first is smart test selection, the second a folder on a schedule, the third a collection running as a monitor.

What is the difference between regression testing and retesting?

Retesting checks that one fixed defect is fixed: the case that failed is run again and must pass. Regression testing checks that the fix broke nothing else: the suite around it is run again and must still pass. In Dokimore retesting is a rerun of the one recorded flow; regression testing is the folder run, or the selection that covers the change.

Smoke, sanity and regression tests: what is the difference?

A smoke test asks whether the build is worth testing at all: it starts, and the main path works. A sanity test asks whether the area that changed works. A regression test asks whether anything else broke. Smoke runs first and stays small, sanity is quick and targeted, and regression is the broad run, which is why choosing the flows a change can reach matters most there.

How do you keep a regression suite from going flaky?

Assert on outcomes rather than on timing, so a slower response does not fail a working feature. Keep more than one way to find each element, captured when the test was recorded, so a renamed id does not fail the step. Read every rescue, because a locator that matched something else is a false pass. And when a test fails without a code change, find the cause before it runs again: a suite rerun until it is green has stopped being evidence.

Try it on your own application

The Free plan asks for no card and runs for 14 days after the trial. Or bring one flow from your product and we record it live.

  • Free plan, no card
  • 14-day Solo Pro trial
  • Demo replies within one business day
  • Data hosted in India