Test a flow before your members see it

A test run plays your flow against your real server, as a member you pick. You get a list of what each step would have done, and nothing is posted.

Test run

Staff applications

Run as

mira

Channel

#applications

Message content

I'd like to join the Crew.

  1. Check roles

    mira isn't in the Crew yet, so the flow carries on.

  2. Save to stored data

    Would save mira's application with today's date.

  3. Send message

    Would post the application in #crew-review for your staff.

  4. Give role

    Would give mira the Applicant role.

Simulated: no real actions were performed.

#crew-review during the test

crew-review
KestrelYesterday at 21:40
quiet night, no new applications
Would be posted here
LanternAPP
New application from mira
I'd like to join the Crew.
miraMemberApplicant · would be given

Nothing was sent.

No message in #crew-review and no new role for mira.

Illustration. The real interface may look different.

mira's Crew application, run as a test. The window opens from the flow editor.

A test starts the way a live run would

You choose a member, a channel and the message they send. The flow reads that member's actual roles and your actual channels, so someone without the Crew role fails the same check she would fail live.

Open any step to see the values it worked with. Test runs don't count towards your server's activity limits.

Who
mira
Where
#applications
Saying
“I'd like to join the Crew.”
A test run in the flow editor with every step marked as passed and a note that nothing was sent.

Screenshot from our test server. Your dashboard may look different.

The same window in the dashboard

A scheduled flow after a test run. Four steps with their status and duration, under the notice that nothing was performed.

The run history tells you why a run failed

Every live run is listed with its status and how long it took. A failed one comes with the reason in a full sentence, like "The bot is missing a permission it needed".

Skipped runs are listed as well, for example when a cooldown was still active.

Run history

Lantern Lounge

FlowStatus
Daily questionSuccess
Welcome kitFailed

Failed

The bot is missing a permission it needed.

Lantern needs Manage Roles, and its role has to sit above Newcomer.

Staff applicationsSkipped
Skipped: cooldown still active.
Staff applicationsSuccess

Illustration. The real interface may look different.

Live runs don't record every step

Only test runs keep the whole trace. From a live run you get the result and, if it failed, the reason, without the steps in between.

When a flow fails live and the reason alone doesn't explain it, run it again as a test with the same member and channel.

Test run

  1. Check roles
  2. Save to stored data
  3. Send message
  4. Give role

All four steps recorded

Live run

  1. Not recorded
  2. Not recorded
  3. Not recorded
  4. Give role

Failed, with the reason

Your AI assistant can run the test for you

An assistant connected to your server tests what it builds before anything goes live. You can also hand it a flow you built yourself.

How building with AI works

You

Run a test of my Staff applications flow as mira and tell me what each step would do.

Your AI assistant

I ran a test as mira in #applications. Nothing was sent.

The flow checks her roles, saves her application, posts it in #crew-review and gives her the Applicant role. All four steps passed.

It's still a draft, so nothing runs until you switch it on.

Questions about testing

No. A test run posts nothing and changes no roles.

Szeretnéd a szervered a bétában?

A V3 nagy rendszer, és még gyorsan változik, ezért a szerverek egyesével érkeznek, nem egyszerre. Minden jelentkezést ember olvas el, és pár napon belül visszajelzünk.

A bétában lévő szerverek döntik el nagyrészt, mit építünk legközelebb. Amit megépítesz, az a tiéd: a flow-k és a modulok a béta végén is megmaradnak.