Smart Text Scanning — How Peace of Mind Detects Explicit Content (2026)

Peace of Mind scans every text node on every page against 2,200+ curated base terms that expand to over 200,000 variants. Every piece of text passes through a full Unicode normalization pipeline before checking — converting homoglyphs, decoding leetspeak, stripping invisible characters, and resolving emoji substitutions. Combined with fuzzy matching and context-aware intelligence, this is the most thorough text scanning system in any content filter.

Last updated: April 12, 2026

2,200+ Base Terms, 200,000+ Total Variants

Peace of Mind maintains a curated database of over 2,200 base terms compiled over years of real-world testing. Each base term is expanded through the normalization pipeline into dozens or hundreds of variants, producing a total vocabulary of over 200,000 matchable strings. The lookup uses a word-split algorithm with Map.has() for O(1) matching — approximately 100 to 1,000 times faster than iterating through every variant sequentially.

The term database covers explicit content vocabulary, platform-specific terminology, slang, abbreviations, and domain-specific language used to seek harmful content. Terms are continuously updated as new evasion patterns emerge.

The Unicode Normalization Pipeline

Before any text is checked against the term database, it passes through a comprehensive normalization pipeline based on the Unicode Consortium's normalization standards:

  • Homoglyph replacement — Cyrillic, Greek, and other script lookalikes are converted to their Latin equivalents. The Cyrillic "a" (U+0430) looks identical to the Latin "a" (U+0061) but is a different character. Peace of Mind catches both.
  • Leetspeak decoding — Digit and symbol substitutions are converted back: 3 becomes e, @ becomes a, $ becomes s, 0 becomes o. Context-aware symbol leetspeak (!, $, @, +, |) only converts when followed by an alphanumeric character, preventing false positives on punctuation.
  • Emoji resolution — Emoji characters used as letter substitutions are resolved to their text equivalents.
  • Zero-width character stripping — Zero-width spaces, zero-width joiners, zero-width non-joiners, and other invisible Unicode characters are removed. These are commonly inserted between letters to break keyword matching in other filters.
  • Diacritical mark removal — Accented characters are decomposed and the accent marks stripped, mapping accented variants back to base Latin characters.
  • Soft hyphen collapse — Soft hyphens and other invisible word-breaking characters are removed so that words split across invisible boundaries are caught.

This pipeline runs on every text node on every page. It executes in microseconds per node, adding no perceptible delay to page rendering.

Fuzzy Matching for Intentional Misspellings

Even after full Unicode normalization, intentional misspellings remain a common bypass strategy. Peace of Mind handles this with fuzzy matching that catches:

  • Missing letters — Omitting a vowel or consonant does not evade detection.
  • Doubled characters — Repeating letters to disguise a word is caught automatically.
  • Adjacent character swaps — Transposing two neighboring letters (a common intentional typo) is detected.
  • Common phonetic substitutions — Spelling changes that preserve pronunciation are resolved.

The fuzzy matching system checks hundreds of thousands of word variants in under 0.05 milliseconds. Combined with Unicode normalization, this makes Peace of Mind's text scanning the most comprehensive of any content filter available. Every creative spelling attempt, every intentional typo, every Unicode trick — all caught silently and automatically.

Context-Aware Intelligence

Most filters treat every website the same. A word is either blocked everywhere or blocked nowhere. This forces a painful choice: cripple your productivity or leave gaps in your protection.

Peace of Mind understands context. Writing about your recovery in Notion? Not blocked. Processing your triggers in Google Docs? Not blocked. Journaling in Standard Notes or Evernote? Not blocked. The same words that are perfectly safe in a recovery journal are caught and blocked instantly on Instagram, Discord, Reddit, and every other platform where they signal seeking behavior.

This context-awareness is a core differentiator. No other content filter distinguishes between therapeutic writing and harmful seeking based on the platform you are using. Peace of Mind treats your recovery writing as sacred while maintaining strict protection everywhere else.

Tiered Response: Soft Block vs Hard Redirect

A single concerning word in a page of otherwise safe content should not lock you out entirely. Peace of Mind uses a graduated response system:

  • Soft block (1-2 unique words) — All media on the page is blurred, all videos are paused, and a banner appears. But the page stays usable. You can still read your email, use your chat app, or browse the thread. Soft blocks do not count toward cool-off escalation.
  • Hard redirect (3+ unique words) — The page is fully redirected to the block page. This is triggered when multiple unique blocked words indicate the page is genuinely harmful rather than incidentally containing a single word.
  • Title/heading match — A blocked word in the page title or main heading triggers an immediate hard redirect regardless of word count, because the page's primary topic is harmful content.
  • Input detection — If you type a blocked word in any input field, it always triggers a hard redirect. This is the outgoing message blocking system — seeking behavior is always escalated.

Words are deduplicated — the same word appearing multiple times across scan sweeps does not escalate the count. Only unique blocked words contribute to the tier decision.

Frequently Asked Questions

How does Peace of Mind detect explicit text on a page?

Peace of Mind scans every text node on every page against 2,200+ base terms that expand to over 200,000 variants. All text passes through a Unicode normalization pipeline that converts homoglyphs, decodes leetspeak, strips invisible characters, and resolves emoji substitutions before checking.

Can Peace of Mind detect misspelled explicit words?

Yes. Fuzzy matching catches missing letters, doubled characters, adjacent swaps, and phonetic substitutions. Combined with Unicode normalization, every intentional misspelling and creative respelling is caught. The system checks hundreds of thousands of variants in under 0.05 milliseconds.

Will Peace of Mind block my recovery journal on Notion or Google Docs?

No. Context-aware intelligence recognizes the difference between consuming explicit content and writing about recovery. On Notion, Google Docs, Standard Notes, Evernote, and other productivity apps, your writing is never blocked — even if you mention triggers while processing your experiences.

What is the difference between a soft block and a hard redirect?

One or two unique blocked words trigger a soft block: media is blurred and videos paused, but the page stays usable. Three or more unique words, or a match in the page title, triggers a full redirect to the block page. This graduated response protects you without punishing incidental word encounters.

Text Scanning That Cannot Be Fooled

Start your 7-day free trial. Cancel anytime.

Start Free Trial