How Do You Stop the Same Message Arriving Twice?

How a duplicate is recognised

Each relayed message leaves a fingerprint built from what it contains: its text, and for an attachment its kind and size. When a new message produces the same fingerprint inside the time window, it is dropped instead of relayed.

You choose what goes into the fingerprint. Text alone is right when the same news is posted with different images; including the attachment is right when the picture is the thing being repeated.

The time window

The window is how far back the comparison looks, five minutes by default. It should match how your sources behave rather than being set as high as possible: a window of an hour will also drop a legitimate repeat — a price update that genuinely says the same thing twice — and you will not be told which.

For overlapping news channels, a few minutes is usually enough. For a source that reposts its pinned message every day, no window will help, and a keyword rule is the better tool.

Near-identical is not identical

The comparison is exact. Two channels publishing the same story with a different opening line produce two different fingerprints, and both are relayed. That is a deliberate limit rather than an oversight: a fuzzy comparison that decides two messages are "close enough" would eventually drop something you wanted, silently, and there is no way to notice what never arrived.

If the duplicates differ only by a fixed prefix or a signature line, a find-and-replace rule that removes it will make the fingerprints match, and the duplicate filter will then do its job.

The protection that is always on

Separately from this setting, there is a safety net that cannot be switched off: the same content sent to the same destination within a few seconds is not sent twice. It exists to stop a loop — two rules pointing at each other, or a destination that is also a source — from multiplying messages endlessly, which is the kind of accident that gets a Telegram account limited.

One consequence worth knowing: if two of your rules deliberately relay the same message to the same destination, only one copy arrives.

Which plans include it

Duplicate detection is part of the paid plans, from Smart upwards. The safety net against loops applies to everyone, on every plan. The plan comparison is on the pricing page.