Meet Sniko.
Sniko is Snikus's own moderation bot. It reads every post, comment, and report the moment it's submitted — and if something breaks community guidelines, Sniko removes it immediately. No waiting for a human to be online, no bad content sitting visible in the meantime.
What Sniko actually does
Reads everything, instantly
Every post, comment, and report is scanned the moment it's submitted — not on a schedule, not after a complaint piles up.
Acts on its own
Spam, banned phrases, and explicit content get removed automatically. Sniko doesn't wait for a moderator to be online to do its job.
Handles reports the same way
When you report something, Sniko re-checks it immediately. If it's a clear violation, it's gone right away — not stuck in a queue.
Knows when to ask for help
If a report isn't a clear-cut case, it's the only time it reaches the human moderation queue — for a judgment call Sniko isn't built to make.
How a report gets handled
Content goes up
A post, comment, or reply gets published — or someone taps "Report" on something that shouldn't be there.
Sniko checks it against community guidelines
Banned phrases, explicit content, spam patterns, link flooding, and shouting-in-caps are all checked in a fraction of a second.
Clear violations are removed immediately
No waiting on a human, no grace period. If it's obviously against the rules, it's taken down right then.
Anything uncertain goes to a person
Edge cases land in the human moderation queue, where the Snikus team makes the final call.
Sniko today
Sniko still works entirely on pattern and rule-based text detection — no third-party AI service involved. It's built to get faster and smarter with future versions, but the principle stays the same: bad content shouldn't get to sit around waiting for a person to notice it.
Changelog
- Expanded the banned-phrase list to catch more scam/DM-bait patterns (fake investment offers, "link in bio" spam, prize-claim bait, contact-harvesting phrases).
- Added detection for phone-number-style digit runs, a common contact-harvesting spam pattern.
- Rule-based text scanning on every post, comment, and report — no external AI service.
- Banned-phrase and explicit-content matching, link-flood detection, repeated-character spam, and excessive-caps shouting.
- Immediate auto-removal for clear violations; anything uncertain escalates to the human moderation queue.
