Core capabilities - Testorim

Authoring

Capability	What it means
Plain-English tests	Describe the workflow. Testorim plans the steps.
Negative tests	Mark a test Should fail and it passes only when the app correctly refuses.
Test data tokens	`{{random.email}}`, `{{date.iso}}` and project-defined variables, substituted at run time.
Mid-run prompts	In a run you are watching in the app, a step can stop and ask you for an OTP, a 2FA code, or a CAPTCHA answer.
Saved procedures	Save a run and replay it without planning it again. Passwords it typed are kept encrypted and typed only on the project’s own hosts.
Playwright export	Export a saved procedure as a standalone Playwright spec, with secrets read from environment variables.

Execution

Capability	What it means
Real browser	Every run drives Chromium through Playwright against your live site. On the Team plan, saved procedures can also replay in Firefox and WebKit.
Mobile and repeat	Replay a procedure at a phone viewport, or three times to surface flakiness.
Live screenshots	Frames stream to the browser panel while the run is in progress.
Environments	Run the same test against staging, production, or a PR preview. A default environment applies when a run names none.
Schedules	Cron expressions per procedure, dispatched server-side.
Time limits	Per-run cap from 5 to 120 minutes, for scrapers and background jobs.
Cancel	Stop a run, or a whole batch, and the browser is closed at once.
Concurrency	Runs up to your plan’s concurrent-run limit. When every test browser is busy, a run waits in line and the chat shows its place.

Checks a step can make

Check	Example
Element state	`visible`, `hidden`, `enabled`, `disabled`, `checked`, `selected`, `readonly`
Text	`contains:`, `not contains:`, `text:`, `value:`
Count	`count:3`, `count:>=1` on a named set of elements
Page load	`loaded`, checking the document’s own status, readiness and content; `images loaded`
API call	After a click, assert the request it fired returned `2xx`
Download	Assert the filename, content type or size of a triggered download
Database	A read-only `SELECT` against a connection you configured
Visual diff	Pixel comparison against a saved baseline for that step and viewport
Performance	LCP / FCP / TTFB against a per-procedure budget

See Steps and expected outcomes for the full grammar.

Evidence captured on every run

A final screenshot, and a screenshot at the moment of each failed step
Before/after page snapshots per step: URL, title and visible text
The element state read back off the page for every assertion
Console messages and network requests recorded during each step
An accessibility scan of the finished page
Performance metrics for each page that loaded
A recording and a Playwright trace of the run
A verdict of Passed, Failed or Needs review, decided by the steps

Beyond a single run

Capability	What it means
Flake tracking	Pass rate over 7/30/90 days, with a composite flakiness score per procedure.
Visual baselines	Set a baseline per step and viewport; later runs diff against it.
Performance budgets	Cap LCP, FCP and TTFB per procedure; a breach marks the run’s performance `degraded`.
Form fuzzing	Generate adversarial payloads for every `type` step in a procedure and fire them as a batch. Each variant checks only its own steps.
Role boundary probe	Walk a list of URLs as a privileged and a low-privilege session and report what both can reach.
Run sharing	A revocable public link to one run’s report, with credentials stripped.

Integrations

GitHub: a sticky PR comment with the run results, and a mirror of your own check runs
Slack: alerts to a channel you pick when a run fails or needs review
Email: the same alerts to a list of recipients
CLI and the HTTP API: trigger runs from CI

Some capabilities are gated by plan. See Plans and limits.