Pricing
Protection is free. Conversation is what costs money.
Almost everything Wardnox does to keep a server safe is ordinary computation running on a server we are already paying for. It costs us nothing per message, so we do not charge for it, on any plan, at any volume. What the paid plans sell is the AI.
On every plan
What you get without paying anything.
This is the part worth reading first. A free Wardnox server and a Scale one are protected by the same engine, and neither has a limit on how much it will scan.
Every deterministic detector, on every plan, with no cap on how many messages are scanned
All enforcement: delete, timeout, kick, ban, and the escalation ladder for repeat offenders
Lockdown, emergency stop, and the automatic pull on a rule that starts misfiring
Member verification, with six challenge types
Welcome messages and banner images
Log channels, moderator actions on every logged decision, and a full audit trail
The rule playground, configuration export and import, and one click restore
The plans
Four, and the second one is a fork rather than a step.
Most people who install a security bot do not want a conversational one. Guard is for them. Pro and Scale are for servers that want the bot answering their members too.
Free
Free
A genuinely good security bot, at no cost, forever.
Any server that wants protection and has not decided anything else yet.
Guard
$9/mo
Full protection, no chatbot.
Servers that want a security bot and nothing else. Most of them. This is not a step towards Pro, and staying here forever is a perfectly good outcome.
Pro
$29/mo
Protection, plus a bot that can actually talk.
Communities that want questions answered from their own documentation rather than guessed at.
Scale
$99/mo
The same product, with the ceilings taken off.
Large or busy servers. Scale has no feature Pro lacks: everything is switched on at Pro, and this buys volume.
Side by side
The whole comparison, in one table.
| Free | Guard | Pro | Scale | |
|---|---|---|---|---|
| Price | Free | $9/mo | $29/mo | $99/mo |
| Messages scanned | Unlimited | Unlimited | Unlimited | Unlimited |
| Word / regex / domain rules | 15 / 3 / 10 | 50 / 10 / 50 | 100 / 25 / 100 | Unlimited |
| AI scam checks | A daily allowance | No daily limit | No daily limit | No daily limit |
| AI credits per month | 50 | 50 | 10,000 | 40,000 |
| Knowledge base | Not included | Not included | 25 sources | 200 sources |
| External threat intelligence | Not included | Not included | Included | Included |
| History | 7 days | 30 days | 90 days | 1 year |
Credits
What a credit is, and what spends one.
One credit, roughly one thing the AI does
Answering a member, checking a borderline message, or reading a document into the knowledge base. A reply is about one credit; a scam check is about half of one.
Only the AI spends them
Scanning, blocking, enforcing, verification, welcome messages and logging are never metered, on any plan, at any volume. If it does not call a language model, it does not cost you a credit.
Running out is quiet
The bot stops joining conversations until the next month. Your members are never shown a message about your billing, because your billing is not their business.
Protection is not affected
Every deterministic detector keeps running when the credits are gone. What pauses is the second opinion on borderline messages, never the scanning itself.
Every server starts on a trial of Pro.
Fourteen days of Pro, with no card and nothing to set up, from the moment Wardnox joins your server. Turn conversation on and let your members talk to it, because that is the part no feature list can tell you.
We are adding servers in batches rather than all at once, so the trial begins when yours is added rather than today. Leave your address and we will tell you when that is.