guide · alt accounts
How to tell an account is an alt
A ban closes an account — not a person. The banned player usually returns the next day on a fresh Discord account with no history at all: zero bans, a clean application, a name nobody recognises. The new account may have no history of its own, but it can still share traces with the old one. This guide covers what those traces are, and which of them are strong versus misleading.
First, a correction
A search for “discord ban lookup” usually carries an expectation: that Discord keeps a global ban list, you enter an ID, and you see whether the person is banned. No such list exists. Discord bans are per-server, and Discord exposes no public API that shares them.
What works in practice is a different question: “have other communities already banned this person, and is this account linked to one that was?” The answer to that does not live at Discord but in shared archives where servers pool their own ban logs. That is where Argos’s ban records come from — co-operating servers, not Discord.
The signals that give an alt away
From strongest to weakest. The order matters: the further down you go, the more likely a coincidence becomes.
Shared game identifiers
The hardest evidence. If two Discord accounts have been seen with the same FiveM license, Steam ID or Xbox Live identity, the same machine and the same game account sit behind them. These are the traces a player leaves on connect; opening a new Discord account does not change them. Beating this requires genuinely replacing the Steam account too — about as expensive as buying the cheat again.
Account creation timing
Nothing on its own, but a lot next to a link. An account created a few days after the banned one was banned, sharing the same identifier, makes the coincidence argument very weak. The Discord user ID itself encodes the creation date; you do not have to take the account’s word for it.
Overlapping server memberships
Two accounts present in the same servers, especially small niche ones. Overlap in large public servers means nothing — millions are there. What counts is appearing together in low-member cheat communities or the same RP server.
Repeated avatars
People often re-upload the same image when they open a new account. Useful, but to be handled with care: a popular meme avatar appears on hundreds of accounts. Only avatars shared by a small number of accounts mean anything.
Name similarity
The weakest link and the biggest producer of false positives. “xNeoX” and “xNeoXx” may be the same person — or two different players who picked the same popular name. Never use name similarity alone; read it only as support for a hard match.
Why naive alt detection links everyone together
The most insidious mistake in alt detection is using signals transitively. A shares an avatar with B, so link them; B shares a name with C, so link them — and now A and C sit in the same cluster having nothing to do with each other. That chain runs away within a few steps.
This is not a theoretical risk. It happened in Argos’s own clustering: with transitive linking through avatar hashes enabled, a single cluster grew to 7,857 accounts — the system was effectively claiming all those people were one person. The cause was one widely used avatar that chained hundreds of unrelated accounts together.
The fix was not to drop the signal but to bound it. Clustering now runs on hard identifiers only; avatar and name similarity no longer form clusters and appear solely as a weaker edge type in the visual graph. An identifier that shows up on an abnormal number of accounts (a shared machine, or spoofed data) is discarded outright, and cluster size is capped. After that change the largest cluster fell from 7,857 to 29.
The reason for telling you this: if an alt-detection tool shows you a long list of “linked accounts”, the right question is not how many it found but what it linked them on. A wrong alt match gets an innocent player banned, and that is hard to notice after the fact.
Where Argos fits — and where it does not
A Discord ID lookup collects these signals on one screen: account age and name history, memberships and customer roles in monitored cheat communities, ban records from co-operating servers, and linked accounts. The Premium identity graph opens those links visually and labels which signal each edge came from — shared identifier, co-membership, avatar or ticket timing. The archive currently spans 400,000+ tracked identities and 110,000+ ban records.
What Argos is NOT
Argos is not Discord’s ban list — no such list exists. It does not see direct messages, unmonitored servers or closed communities, and no query ever shows an IP address. And alt detection produces probability, not certainty: a link is grounds for a decision, not the decision itself. An empty result means “no record”, not “clean”.
How each signal is actually collected
Most tools sold as "alt detection" use one of the same four mechanisms. Knowing which one decides how much weight its verdict deserves — and every one has a known limit.
- Identity shared at connect time: the durable keys a server reads from a player (Discord ID, game identifiers). The hardest signal, because it does not change at the player’s will. Its limit: it only sees someone who reuses the same key — a brand-new account leaves nothing behind.
- Network and device traces: IP address, browser fingerprint, device identifier. One house, one school or one VPN exit links two different people. Its limit: mobile carriers and VPNs share a single IP across thousands of users, so used alone this has the highest false-positive rate of the four.
- Behaviour and timing: accounts created together, moving through the same channels, opening the same tickets. Its limit: friend groups, shared events and teams joining a new server in bulk produce this pattern by themselves.
- Community record: whether that account appears elsewhere with a ban, a customer role or membership in a seller network. Its limit: it covers only communities that share their records; a player banned on a server that shares nothing is invisible here.
Which of these we use, and which we do not
Argos uses two of the four: identity shared at connect time, and community record. Both rest on the account’s history rather than anything the player claims about themselves.
- We use — shared identity: the same Discord ID or game identifier appearing on two accounts. A hard match; action decisions come from it.
- We use — community record: ban records, cheat-community memberships, customer and seller roles. Shown with their source, date and reason.
- We do not use — network and device traces: an IP address is never shown in a query result and is never used for matching. Linking two siblings who share a connection costs more than the cheater you catch with it.
- We do not use — behaviour and timing: these provide context but do not create a match. A moderator’s observation that "these two move together" cannot be the basis of an automated decision.
That means we do not use the signal competitors lead with, and we are not hiding it. What we get in return is a low false-positive rate: name similarity and same-network links do not produce action by default — only recorded matches do.
Frequently asked
Is there a global Discord ban list?
No. Discord bans are per-server, and Discord exposes no public API that shares them; a “global Discord ban list” does not exist. Argos’s ban records come from the ban logs of co-operating servers, not from Discord. So a lookup answers “has one of the servers in this archive banned them?” — not “has Discord banned them?”.
Is alt-account detection conclusive?
No, it is probabilistic. Even the strongest signal — two accounts sharing the same FiveM license — could in theory come from a shared computer. The right use is not to treat a single match as proof but to stack signals: the same identifier, plus an account created after the banned one was banned, plus membership of the same servers, carries weight none of them carries alone.
If the avatar matches, is it the same person?
Usually not. Repeated avatars are a useful but weak signal: a popular meme avatar is used by hundreds of accounts. Argos only considers an avatar link when few accounts share that image; widely used avatars are discarded entirely. An avatar never forms a cluster on its own — it only carries meaning next to a hard identifier match.
Does changing username break the trail?
No. A Discord username and display name change, but the user ID is fixed for the life of the account — Argos keys records to the ID, not the name, and keeps name history separately. If the person opens a completely new account they get a new ID; that is exactly where alt detection comes in, because the new account may have no history of its own but can still share traces with the old one.
Will the person know they were looked up?
No. The lookup runs against Argos’s own archive; no request touches their account, you do not need a server in common, and nobody is notified.
Is alt detection a Premium feature?
A Discord ID lookup is free — 5 per day on the free tier — and shows the number of linked accounts. The breakdown of those links (which account, connected by which signal) and the visual identity graph are Premium features.
Does Discord detect alt accounts itself?
Discord looks at account behaviour: accounts from the same device or IP, newly created accounts and bulk joins can be flagged, and action can follow under its own rules. But there is no public list or API of those signals — a server owner cannot see Discord’s internal flags.
Can a Discord server detect alt accounts?
Not directly — Discord does not expose alt-account information to server owners. Indirectly, yes: your server can run its own check using data it collects — members’ game identifiers, join times, shared servers and message history. Argos Sentry does that automatically as a member joins.
Does Discord ban accounts for being alts?
Discord does not define "being an alt" as a violation; action depends on what the alt does. Bans follow a rule breach such as ban evasion, fraud or spam. On a server, an alt is usually treated as a banned member returning, and handled under that server’s own rules.
How do I tell if two Discord accounts are the same person?
Stack the signals: a shared game identifier is the hardest evidence; accounts created in the same period, overlapping server history and a rare repeated avatar support it. A single signal — especially a name similarity or a popular avatar — is not enough.
What data does alt detection actually use?
Four kinds. Shared durable identifiers read at connect time. Network and device traces such as IP or browser fingerprint. Behaviour and timing patterns. Community records — bans, customer roles, memberships. Each carries a different false-positive rate, and a tool that does not say which one produced a link has not told you enough to act on.
Can a bot reliably detect alt accounts by IP address?
No, and treating an IP match as proof is the most common way to get this wrong. Carrier-grade NAT and VPN exits put thousands of unrelated people behind one address, so an IP match is a lead, never a verdict. It is also why we do not use it for matching — the tool with the highest raw hit rate is usually the one with the most false positives.
Why do accounts from one household keep getting flagged as alts?
Because same-household signals are what naive matching leans on: one IP, one device family, sometimes one browser fingerprint. Two siblings are not alts, and no amount of signal stacking makes them so. The defence is to require a recorded match — a shared durable identifier or a community record — before any automatic action, and to keep an appeal path for everything else.
Related: FiveM player lookup · Pre-whitelist vetting · Argos Sentry — Discord join screening · Banned on other servers?