सामग्री पर जाएँ

यह पृष्ठ अभी आपकी भाषा में उपलब्ध नहीं है। आप अंग्रेज़ी पाठ पढ़ रहे हैं।

Knowledge base

No connection to the server and the offline period has ended. Sign in again.

The app works for one day without reaching the server, and that day has run out. Nothing on your disk is touched: what changes is that new actions wait for a single check with the server to go through.

Where the day starts counting

The clock does not start when the cable comes out. It starts at the last answer the server gave, and while you are signed in the app asks for one every five minutes.

So the day really begins within a few minutes of the connection going away. A short drop in the middle of working hours costs nothing at all.

Twenty four hours and no more

The period is one day exactly. At twenty three hours and fifty nine minutes everything opens as usual, and at twenty four it does not.

The pricing page names the same hours, because the app and the site read that number out of one place.

Nothing is deleted and nothing closes

Profiles, cookies and logins stay exactly where they were. The browsers standing open on screen at that moment keep running until you close them yourself.

The same line in two lengths

The line comes in two lengths, and you may have met either. The strip above the profile list, the account screen and the notification all show it without the last sentence.

The longer one ends with Sign in again and comes from a profile you tried to create just as the period ran out. Both name one cause and both end in the same place.

What waits for the server

Six things wait for the server. The reason stands in a strip above the profile list.

  • Opening a profile

    The button on the card dims and names the reason.

  • Creating and editing profiles

    Import from a table comes through the same door.

  • Flow runs and profile warm-up

    A run started by hand ends before it opens a window.

  • Schedules

    A slot that comes due is passed over.

  • Proxies across many profiles

    Assigning from the library and rotating over a pool.

  • The local API

    A start asked for by your own script gets the same answer.

What still works

The window is not dead. Everything below carries on, and a profile edit is not on this list because it waits with the ones above.

  • The profile list

    Names, groups, notes and proxies all read as usual.

  • Export

    Settings of many profiles, or one profile with its session.

  • Closing a browser

    A window opened earlier closes the normal way.

  • Deleting a profile

    The row goes to the recycle bin as it always does.

  • Checking a proxy

    It goes out through the proxy to a public IP service and never to us.

  • Proxies, flows and groups

    A new row, a new flow or a new group saves as usual.

Schedules are passed over and named

A schedule that comes due under this lock is passed over, and a message says so at the time.

Missed slots are not stored up. A night of warm-up that did not happen does not arrive all at once when the lock lifts.

The way out is a connection

While you stay signed in the app keeps asking every five minutes, and the first answer it receives lifts the lock on its own.

Closing the app and opening it again changes nothing here. The lock is written down and read back at the next start, and signing in again needs the server too, so it helps only once there is a connection to use. Any network will do.

A clock wound backwards ends it sooner

Putting the machine clock back leaves the last check sitting in the future, and the app reads that as a period already over rather than an endless one.

What this article does not cover

  • The other locks. An inactive subscription, a device limit and an expired session each carry their own line and their own way out.
  • What travels to the server during a check. The security page walks through that.
  • The daily pace of actions inside a profile. That is counted whatever the connection does, and this line is not about it.
  • Running with no account at all. A signed out copy never meets this line and works on the free plan instead.

Where this is decided

Each line names where this is decided. Four of them are stands that run and one names code with no stand.

One day without the server, counted from the last successful check
apps/desktop/test/session-restore.js
A clock wound backwards ends the period instead of extending it
apps/desktop/test/session-restore.js
The hours the app counts and the hours the pricing page names are one number
apps/desktop/src/lib/license.js · apps/web/components/pricing/PricingTable.tsx
A successful check with the server lifts the lock on its own
apps/desktop/test/seat-lock-clears.js
A schedule that comes due under a lock is passed over and named
apps/desktop/test/schedule-runs.js

Questions

Does a warning come before the day runs out?
No. One notification arrives at the moment the lock falls, carrying this same reason, and the account screen repeats it.
Does the count start when the connection drops?
It starts at the last answer from the server, which is at most five minutes earlier while the app is running.
Do the browsers already open close?
No. They keep their pages and their sessions, and they close when you close them.
Will restarting the app clear it?
No. The lock is saved and read back on the next start, so the app opens in the same state it was in.
Can the period be made longer?
Not from the app. The same hours apply on every plan, and the pricing page names them.

Still stuck

Support reads the report that travels with your message and answers with the cause, not with a checklist.