How machine-like does this read?
Explore a sample score from 0 to 100 and the phrases behind it. It describes register, not authorship.
300 words per runNo accountNothing stored
Sample text206 chars · 34 words
The score reads the text. It does not identify a model or prove authorship.Open in the editor
This tool is not connected yet. The result below is the published example, run on the sample above. The editor runs the same rules on your own draft.
Result
39% human · 6 matches
Human score39%
MachineHuman
Stock opener
1 match
Filler phrase
1 match
Cliché
2 matches
Vague claim
1 match
Hedge
1 match
How it works
Every phrase match counts against the text, weighted by how often it appears in generated writing and normalised by length. The categories below show what pulled the score down, so a low score always comes with a reason you can act on.
This is not a detector and cannot prove who or what wrote a passage. Human writing that leans on stock business phrasing will score low, which is the point.
Need more than 300 words?
The full editor keeps every revision, explains each change, and polishes 5,000 words a month free.