SafeWatch · Content Moderation Dashboard

Content Analysis Tier 0 · Tier 1 · Tier 2

User Post / Message

Load a sample that exercises Tier 0, Tier 1, or Tier 2.

Choose the language mode for this check.

0 characters

Analysis Result —

Category

—

Verdict

—

Detection

—

Confidence

—

Model Used

—

Tier 0

—

Tier 1

—

Tier 2

—

Analyzed Post

—

Session Stats

Analyzed

Flagged

Safe

—

Flag Rate

Recent Activity 0 entries

No analyses yet this session

Moderation Flow Regex · KoalaAI · Qwen

Tier 0 blocks explicit matches instantly.
Tier 1 scores the content with KoalaAI.
Tier 2 only runs on ambiguous Tier 1 scores.

Tier 0 · Regex + hard flags

Fast rules, Greek direct matches, protective override~1 ms

Tier 1 · KoalaAI/Text-Moderation

Multi-label moderation model~20 ms

Tier 2 · Qwen/Monitor

Optional escalation for ambiguous scoresrare

Ambiguous band 0.30 - 0.70