Chat moderation for games
Keep the peace in game chat.
Banter stays. Harassment, PII and grooming risk don't. Paxmod reads every player message in under 500ms and steps in only when it should, through one API call for Unity, Roblox or any live chat.
- From $0.25 per 1,000 messages
- Fail-open by design
- Layers on Roblox TextFilter
- No sales call
- n
nova_77
gg wp, close one
- m
moth
ur aim is so bad lol
- x
xX_rex_Xx
Message heldEscalatedgrooming risk - k
kiwi
rematch? 1v1 me
- l
lilbean
lets gooo that clutch
The problem
Blocklists can't tell banter from harm
Game chat is fast, coded and full of in-jokes. Blocklists either miss what matters or ban your most engaged players.
Player says
ur so bad lol, uninstall
Banter isn't abuse
- Blocklist
- Blockeda false ban
- Paxmod
- Passedfriendly trash talk
Player says
u seem mature for ur age, whats ur snap?
Grooming uses friendly words
- Blocklist
- Allowedno bad words
- Paxmod
- Escalatedgrooming risk
Player says
f u c k off n00b
Blocklists get gamed
- Blocklist
- Allowedspacing slips past
- Paxmod
- Mutedprofanity, harassment
The console
Every message scored. Every call explained.
Watch decisions land in real time, see where messages go, and pull the gray areas into a queue for your mods.
Messages scored today
184,203
Paxmod acted
10,920
To review
873
Where messages go
- Passed93.6%
- Paxmod acted5.9%
- Needs review0.5%
Live feed
last decision 280ms
- nPASSED280ms
nova_77 / Clean
gg wp everyone
- tMUTED333ms
trashpanda / Harassment
ur f*cking tr4sh, uninstall
- lPASSED386ms
lilbean / Clean
anyone want to squad up?
- xESCALATED439ms
xX_rex_Xx / Grooming risk
u seem mature for 12, add me on snap
- mPASSED492ms
moth / Banter
ur aim is so bad lol
Flags by category, last hour
In production
Working quietly in the chat hot path
Over a quarter of a billion messages moderated, and more than a million every day.
- Messages moderated
- 0M+
- Every day
- 0M+
- Decision time
- <0ms
- Languages
- 0+
- Uptime
- 0.0%
How it works
Send, read, act
- Step 1 of 3
Send
Post each message to one REST endpoint, or drop in the Unity or Roblox SDK. Staging keys are included, and most studios ship in under two weeks.
- Step 2 of 3
Read
Models trained on game chat tell slang from harassment, see through leet-speak, and spot PII and grooming patterns. Add your own rules in plain English.
- Step 3 of 3
Act
Pass, flag, mute, ban or escalate automatically, in under 500ms. Every decision is logged, so you can always show your work.
Try it
Throw your worst message at it
This is the real API in standard mode. Type anything or pick an example, then move the threshold to see how strictness changes the call.
Verdict
Score a message to see it here
- Sexual0%
- Hate and Discrimination0%
- Violence0%
- Dangerous and Criminal Content0%
- Self-Harm0%
- Medical Advice0%
For Roblox developers
Keep TextFilter. Cover what it misses.
Roblox requires TextFilter, and Paxmod doesn't replace it. We layer on top to add grooming-risk detection, PII flagging and your own rules. If our API is ever unreachable, chat fails open and the game keeps going.
Platform
Built for small teams running big communities
Everything live-ops and engineering need, without a sales call or a six-month rollout.
Real-time API
Decisions in under 500ms, right in the chat hot path. If we're ever unreachable, chat fails open so matches never stall.
Custom rules engine
Write rules in code or the visual editor. Combine AI scores with player reputation and context.
Human review queue
Auto-resolve the obvious and send the gray areas to your one or two person mod team.
Analytics & reporting
Trends, dashboards and exportable reports, so you always know what's happening in your community.
Security & compliance
GDPR-ready, with data encrypted in transit and at rest and EU or UK data residency.
Pricing
From $0.25
per 1,000 messages. Free tier included.
See pricing
Banter stays. Harm doesn't.
Start with 1,000 free messages. No credit card, no sales call.