How to use this text tool
- Paste text, upload a text file, or load the sample.
- Keep enough surrounding context for a meaningful review.
- Run the tool and inspect each finding against the source.
- Copy or download the result for editing or human review.
Harassment Phrase Highlighter
Locate common insult, humiliation, exclusion, and coercion patterns and return line-and-column annotations for a moderator to assess in context. Everything runs locally in your browser, and the result stays editable and exportable. Use the findings as a practical first pass, then check the original wording and surrounding context before publishing or taking action.
Text Input
Browser-based processingAd space
What this tool does
The output turns the source into a structured, auditable result instead of silently changing or labeling it. Locations, categories, or missing fields make it easier to revisit the original text.
Deterministic JavaScript rules process the text in the browser. No external model or hidden server decision is used.
A match is a review signal, not proof of intent, policy violation, quality, danger, or legality.
Example
Load the sample and run the tool. The output will show a reproducible highlighter result with counts and, when applicable, source locations.
Use cases
- Pre-publication editorial checks
- Moderation queue triage
- Consistent handoff notes
- Browser-only review of sensitive drafts
Tips for better output
- Keep the original text for comparison.
- Include complete sentences around flagged language.
- Check names, figures, quotations, and policy terms manually.
- Do not treat a zero-result scan as a guarantee.
- Escalate urgent safety concerns under your organization’s procedure.
Processing details
The page uses explicit string parsing and regular-expression patterns. Structured exports contain only the findings produced in the current browser session.
Coded language, spelling variation, slang, sarcasm, quotation, multilingual text, and changing context can cause false positives or false negatives.
Frequently asked questions
Does the Harassment Phrase Highlighter send text to a server?
No. The page processes the text in your browser and does not require an account or external API.
Can this tool make a final editorial or moderation decision?
No. Pattern-based results are review aids. Meaning depends on context, quotation, audience, intent, and local policy.
What input format works best?
Plain UTF-8 text works best. You can type, paste, upload a text-based file, or load the included sample.
Why might the tool miss or over-flag content?
Language is flexible. Spelling variants, coded wording, sarcasm, reclaimed terms, and quoted material can defeat or confuse deterministic patterns.
Can I export the result?
Yes. You can copy the output or download TXT, CSV, or JSON when structured findings are available.
Related text tools
Review guide
| Signal | Meaning |
|---|---|
| Finding | Needs contextual review |
| No match | Not a guarantee |