Tutorial

Run a manual test with the extension

Test Run mode records replay, console, network and a screenshot per step while you work — the same evidence a CI run produces.

Applies toPlans:FreeProBusinessEnterpriseRoles:OwnerAdminDeveloperQaViewerWhere:ExtensionWeb app

In most test tools, a manual pass is a status and a note. Nobody can re-examine it later, so a disputed result comes down to somebody's memory. Test Run mode exists so that a human execution leaves the same kind of proof a machine execution leaves.

Before you start

You need the BugMojo extension installed (install the extension) and at least one case assigned to you. If you were sent a signed tester link instead of an account, open that link — it carries your assignments and needs no sign-in.

Execute a case

Open the tab you want to test on, then open the BugMojo side panel and choose Test runs.
Pick the assigned case. Its preconditions and steps appear in the panel, in order.
Choose Start. A recording border appears around the tab: recording is running and the overlay shows the current step.
Do the first step in the page. Then mark it Pass, Fail or Blocked in the panel — the extension takes a screenshot at that moment and attaches it to the step.
Add a note on any step where the result is not obvious. Notes are what a reader needs when a screenshot alone is ambiguous.
Work through the remaining steps. Navigating, opening a new page in the same tab, or closing and reopening the panel is all fine.
Choose Submit. The extension assembles the artifacts, uploads them, and writes the result against the exact case revision you executed.

What is captured

Artifact	What it is
Session replay	An rrweb DOM recording of the whole run, scrubbable in the dashboard.
Console log	Console output captured across every page in the session, de-duplicated.
Network log	Requests with status and timing, de-duplicated across page loads.
Screenshots	One per step, taken at the moment you marked it.
Step log	Each step's verdict, note and timestamp, rolled up into the overall result.

Replay chunks are spliced, so they are concatenated in order and never de-duplicated — rrweb events carry no id, and dropping "duplicate" events would break the replay. Console and network are ring-buffer snapshots, so they overlap heavily and are de-duplicated by entry identity. The result is one clean artifact of each kind.

The run closes when the queue empties

Submitting a result finds or creates the open manual run for that plan version and stamps it back onto the assignment, so the queue and the run agree from then on. When the last open assignment in the run is finished, the run completes on its own. A manual run ends when the work is done, not on a timer.

If something goes wrong

The panel says an artifact was dropped
Likely cause: One artifact exceeded the per-evidence size cap, or the session hit the per-result artifact limit.
The rest of the run is unaffected and still submits. Keep manual runs focused on one case rather than wandering the whole app, and the caps are rarely reached.
Recording stopped when you switched tabs
Likely cause: Test Run mode is pinned to the tab you started it on.
Return to that tab and continue. If the flow genuinely spans tabs, mark the step blocked and note it — a case that needs two tabs usually needs rewriting.
The replay looks blank in the dashboard
Likely cause: Canvas, WebGL or cross-origin iframe content does not survive DOM replay.
The per-step screenshots still show what happened. See replay is blank or missing events.
You submitted the wrong verdict
Likely cause: A step was marked before the page settled.
Re-run the case. Results are per run, so a second execution records a new result rather than overwriting the first — and both stay in the history.

Common questions

What happens if I navigate to another page mid-test?

Nothing is lost. The session and its capture chunks live in the extension's local database, so a navigation, a closed panel or an evicted service worker does not end the run.

Can I run two cases at once?

No — one active session at a time. Two concurrent sessions would race over the same recorder, and a tester works one case at a time anyway.

Does redaction still apply during a test run?

Yes. The same client-side masking runs before anything is uploaded, and an organisation-enforced redaction floor cannot be loosened by a tester.

What decides the overall verdict?

The per-step verdicts roll up. Any failed step fails the case; blocked steps mark it blocked; all passed marks it passed.

Is a manual result really the same as a CI result?

Yes — same evidence path, same hashing, same size caps, same provenance columns. That parity is deliberate, and it is what the client report relies on.

Was this helpful?

Updated Sep 19, 2026 · Still stuck? Contact support