Kind Whisper Safe
Got a message that doesn't feel right?
Paste it here. Safe points out language that may be threatening, insulting, or pressuring, explains why, and suggests what you could do next.
Possible harm, explained in plain words.
Safe checks a message for common signs of online harm and tells you what it found and why that might matter. It describes the message, never the person who sent it.
Threats and intimidation
Threats of violence, “I know where you live,” and messages meant to scare.
Pushing self-harm
Messages telling someone to hurt themselves or that they'd be better off gone.
Insults and degrading language
Name-calling and put-downs aimed at the reader.
Sexual pressure
Unwanted requests for photos and threats to share images.
Pressure and secrecy
“Don't tell anyone,” “if you loved me you would,” and other manipulation.
Hateful language
Common phrases attacking people for who they are.
Doxxing
Someone's address, phone, or ID number together with a threat or a plan to post it.
Repeated messages
In a pasted conversation, several concerning messages from the same sender. One message alone is never treated as a pattern.
How to read the result
No clear concern detected
Nothing Safe recognizes. Not proof that the message is fine.
Potential concern
Language worth a second look, including insults, pressure, and anything quoted or marked as a joke.
Serious concern to review
Direct threats, pushing self-harm, sexual pressure, or doxxing. Get support from someone you trust.
How it works
Transparent rules, on your device.
Safe compares the message with phrase lists for each supported language and notes when words are quoted, reported (“he said…”), or followed by a joke marker. It skips common idioms like “you're killing it.” It runs in your browser, doesn't use an AI model, and doesn't send or save your text. It can't understand context the way a person can.
Who it's for
- Teens and young people who got a message that doesn't feel right.
- Parents and caregivers helping a child decide what to do.
- Teachers, counselors, and youth workers looking at a message a student shared.
- Anyone who wants a calm second look before replying.
Limits
Information, not a verdict.
- It can miss harmful messages, especially new slang, sarcasm, images, voice notes, and harassment spread over time.
- It can flag harmless messages, like friends teasing each other.
- It doesn't know who sent the message, what happened before, or what the sender meant.
- It never contacts anyone, files reports, or labels a person.
How well Safe covers each language.
Full: wide phrase lists including slang. Limited: direct threats, pushing self-harm, common insults, sexual pressure, and secrecy demands only. A clean result means less in a limited language.
| Language | Language support |
|---|---|
| English | Full |
| Tiếng Việt | Full |
| Deutsch | Limited |
| Español | Limited |
| Français | Limited |
| Português | Limited |
| Italiano | Limited |
| Nederlands | Limited |
| 日本語 | Limited |
| 한국어 | Limited |
| 简体中文 | Limited |
| 繁體中文 | Limited |
Questions
Does Safe send the message anywhere?
No. It checks the message in your browser and doesn't save it. Pressing Clear removes it from the page.
Will it report the sender?
No. Safe never contacts anyone or files reports. If you want to report, use the platform's reporting tools.
Does one bad message mean someone is a bully?
No. Bullying is usually repeated. Safe describes what a message may contain, never who a person is.
Is it free?
Yes, with no account and no daily limit.