Unofficial project — not affiliated with blader/humanizer or any AI vendor.
AI writing patterns

The 35-pattern review checklist.

The official Humanizer skill reviews 35 broad writing patterns. This independent browser checker can flag 12 concrete text signals automatically; the other patterns need context and human judgment.

Browser checks

12 signals the page can flag automatically

A match is a review prompt, not proof that AI wrote the text.

Punctuation

weight ·13
Punctuationvery common

Em-dash overuse

Long dashes dropped mid-sentence where a comma or full stop belongs. Density is the giveaway.

Why AI overuses the em dash

It’s fast — really fast — and simple.
It’s fast, really fast, and simple.

Diction

weight ·11
Dictionliteral check

“Delve” & friends

Delve, tapestry, realm, leverage, robust, navigate, underscore, foster, garner. Model house-style.

Let’s delve into this rich tapestry.
Let’s look at this.
Dictioncommon

Hype clichés

“Game-changer”, “ever-evolving”, “state-of-the-art”, “next-level” — marketing filler at scale.

A game-changing, next-level tool.
A useful tool.

Structure

weight ·12–15
Structurevery common

The rule of three

A compulsive reach for triads: “simple, powerful, and elegant.” One is rhetoric; a cluster is a tell.

Fast, simple, and reliable.
Fast and reliable.
Structurecommon

“It’s not just X, it’s Y”

Negation followed by elevation can sound formulaic when the contrast adds no concrete meaning.

It’s not just a tool — it’s a movement.
It’s a useful tool.
Structurecommon

“Not only… but also”

A balanced correlative the model leans on to sound emphatic and complete.

Not only fast but also secure.
Fast and secure.
Structurecommon

“Whether you’re… or…”

Audience-spanning setups — “whether you’re a beginner or an expert” — a marketing-copy reflex.

Whether you’re new or seasoned…
(name the actual reader)

Filler

weight ·12–14
Fillercommon

Hedged meta-commentary

“It’s worth noting”, “it’s important to note” — throat-clearing that announces a point instead of making it.

It’s worth noting that speed matters.
Speed matters.
Fillercommon

Grand openers

“In today’s fast-paced world…”, “In the realm of…”. Scene-setting that delays the real sentence.

In today’s fast-paced world, speed wins.
Speed wins.
Fillercommon

“Let’s dive in”

“Dive into”, “dive deep”, “let’s dive in” — filler enthusiasm that rarely earns its place.

Let’s dive deep into the details.
Here are the details.

Transition

weight ·8–11
Transitioncommon

Stacked formal transitions

“Moreover,” “Furthermore,” “Additionally,” back-to-back — a model padding for flow.

Moreover, it’s fast. Furthermore…
It’s also fast.
Transitioncommon

Signposted conclusions

“In conclusion,” “at the end of the day,” “all in all” — explicit wrap-up markers humans drop.

In conclusion, it’s worth trying.
It’s worth trying.
Manual review example

Signals that need context

The browser checker does not flag this example automatically. Decide whether the intensifiers add useful meaning before changing them.

Dictionmanual review

Empty intensifiers

Words such as “truly”, “incredibly”, “surprisingly”, and “remarkably” may add emphasis or may only inflate a claim.

It’s truly, incredibly powerful.
Ask what evidence makes it powerful.
Official upstream checklist

All 35 Humanizer patterns

These names follow upstream v2.11.2. Many depend on meaning, sources, formatting, or the writer's own voice, so a simple regular expression cannot judge them safely.

Content patterns

  1. Inflated importance and legacy
  2. Name-dropping to prove importance
  3. Shallow -ing analysis
  4. Sales language
  5. Vague sources
  6. Formulaic challenges and outlook

Language and grammar

  1. Overused AI words
  2. Avoiding is and are
  3. Not X but Y and clipped endings
  4. Forced groups of three
  5. Changing names and repeated openings
  6. False from X to Y ranges
  7. Passive voice and missing subjects

Style patterns, part 1

  1. Em and en dashes
  2. Too much bold text
  3. Lists with bold mini-headings
  4. Title case in headings
  5. Emojis
  6. Curly quotes

Chatbot, filler, and hedging

  1. Chatbot text left in the answer
  2. Knowledge-limit disclaimers and guesses
  3. Overly agreeable tone
  4. Filler phrases
  5. Too many qualifiers
  6. Generic positive endings

Style patterns, part 2

  1. Too many hyphenated word pairs
  2. A fake deeper truth
  3. Announcing the next point
  4. A heading repeated below itself
  5. Writing about the old version
  6. Forced punchlines and fragments
  7. Formulaic sayings
  8. Fake-candid openings
  9. Answering objections no one raised
  10. Rejecting fake alternatives

Source: official Humanizer README at the reviewed commit. Upstream content is MIT licensed; read the site's complete third-party notice.

How scoring works

The Pattern Score, in plain terms

Each detector contributes a weighted amount based on how often its pattern appears, with a sub-linear bonus for repeats and a light correction for very short text. The total is clamped to 0–100. It’s a transparent heuristic — a signal about writing habits, not a verdict on authorship.

0–22few checked signals
23–48some checked signals
49–72many checked signals
73–100dense checked signals

Spot these in your own writing

Paste any text and see which patterns light up.

35 AI Writing Patterns — Humanizer Review Checklist