StreamSafe Phonetic Shield

Scan chat messages, TTS donations, and bait comments for phonetic traps, homoglyphs, invisible zero-width chars, and stumbling baits before reading aloud on stream.

Streamer Safe Reader Preview

● High Risk Bait Trap Detected
Safe Teleprompter (What to glance at before reading) Risk Score: 92/100
2
Phonetic / Homoglyph Traps
0
Zero-Width / Invisible Bytes
High
Acoustic Bait Match
Engine Audit Findings 2 triggers identified
Ready. Scan any comment or donation feed.

Why Streamers Get Trapped on Live Streams

Streamers reading hundreds of fast-scrolling comments or TTS donations inevitably rely on rapid ocular sub-vocalization. Malicious viewers exploit this cognitive shortcut using:

  • Phonetic Homophones: Innocent-looking pairs ("Nick Gurr", "Knee grow", "Bofa D") that sound identical to TOS-violating slurs when spoken quickly.
  • Tongue Twisters: Syllable sequencing designed to cause accidental linguistic slips on stream.
  • Cyrillic & Unicode Cloaking: Visually identical glyphs (e.g. Cyrillic 'а', 'е', 'о', zero-width spaces U+200B) that bypass standard keyword filters.

How StreamSafe Shield Protects You

This browser-local inspection engine runs four defensive passes:

  • Unicode Normalization: Strips invisible spaces, decomposes confusable homoglyphs (NFKD canonical form).
  • Phonetic Token Compaction: Combines adjacent 2-word and 3-word n-grams to test if word boundaries conceal slurs.
  • Acoustic Metaphone Mapping: Matches sounds-alike pronunciation hashes rather than naive spelling.
  • Safe Teleprompter: Highlights verbal hazard points so the streamer pauses before speaking.
Enjoy this tool? Build your own with Super