Comment Toxicity & Content Moderation Checker

Evaluate forum comments, chat messages, and user-generated text for toxicity, severe profanity, personal insults, threats.

Comment Text & Moderation Parameters

Sensitivity Threshold:

Input Comment / User Post

Test Samples:

APPROVED • SAFE TO PUBLISH Overall Toxicity: 4%

This comment passes all moderation criteria. No harassing, threatening, discriminatory, or profane content detected.

Perspective Multi-Attribute Scoring Breakdown

General Toxicity 4%

Severe Toxicity 1%

Insult / Harassment 3%

Profanity & Obscenity 2%

Physical Threat 0%

Identity Attack / Hate 1%

Commercial Spam / Links 5%

API Response Payload (RFC 8259 JSON)

{
  "flagged": false,
  "action": "APPROVE",
  "scores": {
    "toxicity": 0.04,
    "severe_toxicity": 0.01,
    "insult": 0.03,
    "profanity": 0.02,
    "threat": 0.00,
    "identity_attack": 0.01,
    "spam": 0.05
  },
  "evaluated_at": "2026-09-09T10:45:00Z"
}

About this tool

How to use it

Paste user comment, chat post, or forum text into the input area or select a preloaded sample comment.
Choose your community moderation sensitivity threshold (Strict 50%, Standard 70%, or Permissive 85%).
Click 'Analyze Comment' to inspect multi-attribute probability bars across toxicity, severe toxicity, insults, profanity, threats, and spam.
Review automated moderation actions (APPROVE, FLAG FOR REVIEW, AUTO-REJECT) and copy the structured JSON API response.

Why it matters

Online communities and web applications face constant influxes of abusive comments, personal harassment, profanity, and promotional spam. This tool provides automated toxicity scoring and moderation recommendations to protect community trust.

When to use it

Use when building community comment sections, setting up content moderation filters, auditing customer feedback, or testing automated content filtering thresholds.

Common questions

What is the Perspective API moderation standard?

Developed by Jigsaw and Google, Perspective API uses machine learning models to identify toxic language by scoring the probability that a comment will cause someone to leave a conversation.

What do the sensitivity thresholds mean?

A strict threshold (50%) will flag even mildly negative comments for review, while a standard threshold (70%) and permissive threshold (85%) allow more casual banter and only reject unambiguously toxic or threatening speech.

Can I integrate this tool's response schema into my backend?

Yes. The generated JSON output conforms to industry-standard moderation payloads with boolean flagged status, action recommendations, and normalized 0.0-1.0 attribute scores.