0
Members watched
It recognizes the same person even when they switch accounts, devices or networks, and decides what to do. Every decision is saved and can’t be edited or erased. It starts on Discord, but it doesn’t stop there.
A raider joins with a VPN and a two-day-old account, spams six channels, and trips a trap. Watch what your staff would see — and what they would not have to do.
Right now, on any platform, if you get banned you cannot prove what happened — and neither can they. In July, Discord wrongly banned around 8,200 accounts and nobody could show anything either way. Here, every decision comes with a piece of paper you can check yourself.
What actually decides everything — measuring a connection, matching it against what it already knows, reaching a verdict and saving it — does not depend on Discord. Discord only comes in at the end, to apply the decision: give or take a role.
Most security bots do one thing well. Axis does the whole job — and keeps the evidence.
Airplane mode, a router restart, a neighbour’s wifi — the IP changes, the device doesn’t. And the fingerprint is matched PIECE BY PIECE: switching browser or monitor no longer turns you into someone else. If the connection claims one country and the browser says another, no VPN fixes that either. IP (IPv6 prefix), device fingerprint, security cookie, TOR, VPN, proxies and virtual machines.
Mass delete and ban protection, role hardening, and full backups. Guardian mode watches trusted admins too, because that is how servers actually die. And if something breaks, it can be undone: the deleted channels and roles come back, with their permissions. And you can rehearse the recovery whenever you like, without touching a thing: it tells you whether your backup could actually be restored today — and it warns you by itself the day it could not.
Join-wave detection plus behaviour scoring. It acts on what someone does — invites, mass mentions, coordinated copy-paste — never on "you are new and you type fast".
Every deleted or edited message, with a copy of its images, video, audio and stickers. Deleting it does not erase it.
A member dossier: who invited them, how many they brought, sanctions, name history, alt accounts, and everything they have done.
Blocks live only in the server that issued them. Nobody pushes bans into your community from outside, and you can always undo one.
Which invite each member used and who created it. Nickname and username history. Ghost pings. Voice activity. Boosts.
Automatic transcripts as a Discord-style web page, and an appeal system — because automated systems get it wrong sometimes.
Search every event, filter, export, moderate, and read a full dossier. Granular permissions per staff role.
A public profile card anyone can pull up: verified, how long they have been around, and badges they actually earned — Veteran, Clean record, Endorsed, Pioneer. It shows only the good: never warnings, cases or risk. New arrivals get theirs the moment they verify, in front of the whole server.
When a case leaves Discord — harassment, threats, fraud — what do you hand over, forty screenshots? A case file gathers everything about one person, signed, with proof that each event was in the log that day. Anyone can verify it without asking you for anything. No other people’s data, no IPs.
One check is easy to beat. Fifteen, stacked, are not. Each layer catches exactly what the one before it can’t — and the last one isn’t even code.
Every event, searchable. Every member, on one screen. A full dossier: who invited them, their alts, their history — the tools an investigation actually needs.
Ask ONYX, from your own signup form, whether this visit is somebody you already blocked. It answers with a signed verdict and you decide what to do. No Discord account involved anywhere.
Sign up with an email, create a key, and the quickstart fills itself in with that key — in curl, PowerShell, Node or Python. Nothing to install.
curl -H "Authorization: Bearer onx_live_…" \ https://argos.is-a.dev/api/v1/me
POST /api/v1/identity/challengeFrom your backend, with your key. You get back a one-use URL.
GET /check/{nonce}A redirect or an iframe. Connection, browser and device get measured in about a second — and the person is told nothing.
GET /api/v1/identity/verdict/{nonce}Signed with Ed25519. Check it yourself: the public key travels in the same response.
deny means blocked on your site. Never on somebody else’s: what one customer blocks does not travel to the next, because a shared blocklist is one mistake away from banning a stranger everywhere at once.Ask whether other sites have blocked this same person — without telling us who you are asking about, and without anyone’s block ever becoming yours.
Every site starts from zero. Somebody who has been hopping from shop to shop for six months is a stranger at the seventh, even though the other six have them on file.
Now one stranger’s mistake — or one angry admin — shuts a door on you at a site where you have never done anything. And you will never find out why.
At most it moves an allow to review — «take a look». Turning somebody away stays the decision of the site where something actually happened.
Below two, nothing is returned and nothing is counted. One site’s opinion does not exist here — that is what stops a single mistake from travelling.
You send the first three characters of a mark and get back the whole bucket; you find yours on your own machine. Not a promise not to look — we cannot. Same trick Have I Been Pwned uses for passwords.
Unblock somebody and your contribution is withdrawn everywhere. Otherwise a mistake would keep counting against them on other people’s sites forever.
GET /api/v1/identity/quorum/067{ "entries": [
{ "suffix": "045fb…", "sites": 3 },
{ "suffix": "0a91c…", "sites": 2 },
… 14 more in this bucket
] }95 commands. Names and descriptions appear in your own Discord language — and /help explains all of it without leaving the app.
/verificationEntry portal: IP, device, TOR, VPN and captcha/behaviourCatches raiders by what they do, not who they are/antinukeMass delete/ban protection + Guardian mode/antiraidJoin-wave detection/automodSpam, scams, invites, mass mentions, caps/hardenStrips dangerous perms from every role/lockdownLocks the whole server, reopens it exactly as it was/backupFull snapshot: roles, channels and permissions/banipBans someone and every alt on their connection/channelPer-channel rules (images only, no links…)/ban · /unban · /tempbanWith reason, logged and appealable/kick · /mute · /unmuteTimeouts with readable durations/warn · /warningsWarnings with automatic escalation/isolate · /unisolateQuarantine, restoring their roles after/clearBulk delete, with filters/lock · /unlock · /slowmodeChannel control/give-role · /remove-roleRoles you allow, from Discord or the panel/note · /historyStaff notes and full record/saySpeak as the bot (helpers too)/escalationAuto-punishment by number of warnings/userProfile, risk level and known alts/watchlistAlerts you when someone does anything/surveil-allMirrors everything a person does to a staff channel/snipeThe last deleted message in the channel/transcriptChannel history as a Discord-style web page/reportReports with image and video evidence/serverStatus and security posture/panelLink to the web panel/configureRoles, channels, language, log level/verification setupCreates roles, channel and permissions for you/twitchAnnounces when a streamer goes live/syncImports sanctions from another bot/helpEverything, explained inside DiscordYou did not fill in anything. You did not click anything. This is what any site — and therefore the verification portal — can read the moment you arrive.
The sections below read things from your browser — screen, language, time zone, graphics card — and also give your graphics card two jobs and time how long it takes, to show you what any website can know about you without asking.
It never leaves your screen: it is computed right here and is not sent or stored anywhere.
Above, your GPU name is a label: one line in the inspector changes it. Here nothing is asked — the chip is given two jobs and timed. What comes out cannot be edited.
Every dot is a country where Argos stopped an entry attempt. Only the country is counted — never who, never from which server, never an IP.
Having the bot is not the same as having it set up. These communities meet both bars: 5,000 members or more, and 95% of the protection actually enabled.
Commands, logs, the web panel and the verification portal — all of it. Axis picks up your server’s Discord language on its own. Change it any time with /configure language.
Nothing to install, nothing to host. Add it and configure from Discord.
It asks only for the permissions it actually uses — never Administrator. A bot that asks for everything is one you should turn down.
/verification setup mode:webIt creates the roles, the channel and the permissions. Your existing members are untouched — only new arrivals verify.
/verification tor on · /verification fingerprint on · /behaviour enable action:flag
Leave behaviour detection on alert-only for a few days and read the alerts. Switch it to ban once you have seen it be right on your community.
A security bot sees a lot. Here is what it never touches.
Connections, device fingerprints and the security cookie are stored hashed (HMAC-SHA256) and deleted automatically after 90 days.
No files, no location, no DMs, no voice call audio. Public server content only, and only for security.
What is processed, why, for how long, and your rights — in plain language, not legal fog.
The questions a server owner actually asks before adding a security bot.
The Strom Corporation has been building since 2021. It works through three specialist studios under one roof — from games to artificial intelligence to security — that together form a single ecosystem. Axis is the flagship of that security work.
It was not written to fill a feature list. It was written for a real community under attack, by people who had to sit and watch the logs at 3 AM. Every layer exists because something got through once — and we made sure it never would again.
The studio behind Axis. It builds the group’s games and bots, and everything it ships to communities.
Defence against scams, fraud and threats inside communities. Its craft is what shapes how Axis thinks about attackers.
Reach and growth: a marketplace and asset store built to push projects and creators forward.
Where it all started. Original titles built in-house — I.F.C., War Field and many more — on the studio’s own engine.
Artificial intelligence with a life of its own. StromBot and Aegis Bot don’t just answer — they build and run their own living worlds.
The area that guards large communities. Home of Axis — anti-nuke, anti-raid and ban-evasion, battle-tested against real attacks.
Home of Axis — join the studioFree, no premium tier, and every server stays isolated from every other. 59,767 members are already behind it.