This is how Telegram links accounts that never share an address: they are not linked by IP alone. Every session carries a set of device parameters, and if ten accounts share the same set, they read as one group no matter which proxies they connect through.
What the set contains
Device model, operating system, app version, interface language. None of this is secret — the same data is shown to you in the list of active sessions in your own security settings. Open it and look at how your devices are described.
Network data sits alongside it: proxies for Telegram accounts.
Why sameness hurts
Real users are varied: an old phone with an ancient app version here, a new tablet there. A set of ten accounts with byte-identical parameters does not occur in nature, and that is precisely what makes them visible.
Different proxies do not rescue this. The address differs, everything else matches, and the cluster is still apparent.
What to do
- Different values across accounts. Not clever ones — just different.
- Consistency within an account. This matters more than variety between them: one account connecting from a different device every day looks like a compromised account.
- Agreement with everything else. Interface language, phone number region and proxy region that contradict each other are a signal of their own.
What this does not solve
The fingerprint is one component, not a defence. An account with immaculate parameters that sends identical messages to strangers gets restricted on reports, and no configuration prevents that — why Telegram restricts accounts.
The order of importance is behaviour, then reports, then network, then fingerprint. Start at the top — quiet hours.
How Neurogram does this
Everything above is manual work that runs into the number of hours in a day. Neurogram does it on your own accounts and without you: it collects channels and audiences, writes comments and replies with a language model, and keeps limits and quiet hours so the accounts survive. See pricing or look at your case.