Skip to content
QAClan

browser + api regression · runs on your machine

Regression testing for the browser and the API. Recorded once.

QAClan captures your browser flow and the API calls it fires, then replays both as one suite — on your machine, the same way every time.

Get started free
curl -fsSL https://raw.githubusercontent.com/qaclan/agent/master/install.sh | sh

  • Runs on your machine

    ~/.qaclan

  • Deterministic execution

    same steps, every run

  • Works offline

    sync is optional

  • Imports what you have

    HAR · OpenAPI · Postman · Bruno · curl

  • Linux, macOS, Windows

    single binary

  • Source available

    BSL 1.1

the situation

Regression gets cut because it's manual.

every release

  • API calls in one tool, browser flows in another
  • A login token copied by hand between them
  • Two reports to reconcile before sign-off
  • History stranded on one laptop

so

  • First thing cut when a release is late
  • Flaky failures re-run, not fixed
  • “Did we break anything?” answered from memory

the part nobody else does

You don't write the API tests. The journey already fired them.

  1. 01

    Record the flow

    Click through it. Close the browser.

  2. 02

    Capture what it fired

    A second session lists every call, and asks which to keep.

  3. 03

    Assert on what matters

    A field, an operator, a value.

real sessionRecorded against a live app. Every request in the list was fired by the journey.

one ordered run

Both layers. One suite. One report.

browser

You don't write the test.

Click the journey. The script is saved, in order, ready to run again.

Web testing
A suite named Checkout regression in the QAClan agent, holding three scripts, its last run marked failedA suite named Checkout regression in the QAClan agent, holding three scripts, its last run marked failed
one suiteThree scripts, in the order you set.

api

You don't rebuild it elsewhere.

The calls behind that journey become a collection you can assert on.

API testing
The QAClan agent's API workspace: Fulfilment, Orders and Customers collections holding the captured requests, with List orders open on its params, auth, body and assertions tabsThe QAClan agent's API workspace: Fulfilment, Orders and Customers collections holding the captured requests, with List orders open on its params, auth, body and assertions tabs
one collectionParams, auth, body and assertions, per request.

One browser context: a login in step one still holds in step three, so nothing gets copied by hand.

execution model

You don't re-run it to find out if the failure was real.

A test that picks its own steps at run time isn't a regression test — red might just be the tool changing its mind. QAClan runs the steps you recorded, in your order.

AI reads finished results to flag a flaky script. It never touches a step.

  • No self-healing selectors
  • No generated steps at run time
  • No model in the execution path

when it goes red

You don't reproduce it. It's already on disk.

A failed run in the QAClan agent: two scripts passed, Place order failed with ELEMENT_NOT_FOUND, a plain-language cause and a suggested next stepA failed run in the QAClan agent: two scripts passed, Place order failed with ELEMENT_NOT_FOUND, a plain-language cause and a suggested next step
real runThe cause in plain language, and what to do about it.
  • Screenshot at the moment it failed
  • Console and page errors
  • Requests that failed mid-run
  • One HTML file to hand to a developer

agent and server

You don't run a server to start. You add one to share.

How the local agent and the team server compare across 3 dimensions
 AgentfreeServer$30/user/mo
Where it runsYour machine, at localhost:7823qaclan.com, managed by us
What you getRecord, import, organise, run, local reportsHistory across machines, flakiness, trends, AI analysis, seats
How they fit togetherThe execution layer — always requiredThe visibility layer — optional, added when a team needs it
QAClan server reports showing suite health, pass-rate trend and duration trends across a workspaceQAClan server reports showing suite health, pass-rate trend and duration trends across a workspace
six reportsAcross every machine that synced — no dashboard to build.
  • Flakiness index
  • Pass-rate trend
  • Duration trends
  • Failure points
  • Suite health
  • Time to green
All six rows comparedWhat each report answers

pricing

Free for one person. Priced per seat for a team.

There is no trial to start and no clock to beat. Community is the free tier, permanently.

Community

$0 — free forever, no limits

For individual testers.

Get started free

Most popular

Team

$30/user/mo

For QA teams who ship with confidence.

Upgrade to Team

What each plan includes, seat rules and billing questions

getting started

Record your next regression pass instead of repeating it.

Node.js 18+ and Python 3.9+. Nothing else to configure — install guide.

Get started freeRead the docs
curl -fsSL https://raw.githubusercontent.com/qaclan/agent/master/install.sh | sh
  1. 01

    Get your auth key

    Sign in with Google. Copy it from Settings.

  2. 02

    Install the agent

    One command. Its own Playwright and Chromium.

  3. 03

    Record

    Opens at localhost:7823.

    qaclan login --key <your_auth_key> && qaclan serve