How ScanForAI works

How the score works, how to read a result fairly, what AI detection cannot do, and the API. You can use the API to find AI writing and to clean it up. ScanForAI gives an AI-writing score and shows the reasons behind it, one by one. It is a signal, not a verdict.

01

Methodology

Which writing signals we measure: rhythm, word choice, phrasing, how expected the words are, and structure. How we combine them into one score.

Read the method →

02

Interpreting your score

What the bands mean (reads human → reads like AI), why a score is not proof, and how to act on a result responsibly.

How to read it →

03

API

Scan text from your own software. See the addresses to call, what to send and what you get back, the limits, and how to use API keys.

API reference →

04

MCP server

Add ScanForAI to the AI assistant you already use, such as Claude Code, Cursor or VS Code. Then it can check text for signs of AI writing right in the chat.

Set it up →

Key questions

The short answers

What does ScanForAI actually detect?

It looks for patterns that AI text often has. The sentences have a very even rhythm. It uses words that AI likes, such as “delve” or “a testament to”. It uses few short forms like “it’s” or “don’t”. It uses the “not just X, it’s Y” pattern. It leans on lists and headings. Its words are very easy to predict. The scanner measures how often these patterns appear and turns them into one score. Then it shows which patterns it found.

Is it accurate? Can it be wrong?

It’s a strong signal, not proof. No detector — including ScanForAI — is 100% accurate. Formal, technical, or non-native human writing can score high; edited or newer AI can score low. Treat the score as one input, never the sole basis for a decision about a person, and never as grounds to accuse. Lightly editing AI text can lower the score, and the newest “human-tuned” models can evade detection almost entirely — a real, honest limit.

Does it store my text?

Anonymous scans are never stored — the text is processed only to produce the score, then discarded. Signed-in scans are saved only for your own history and can be deleted; see the Privacy Policy.

Does it work on any type of writing?

Best on 300+ words. It reads more reliably on writing than on short snippets, code, or lists. You can tell it the text type (e.g. an academic essay), and it relaxes the “formal register” signals so genuinely formal human writing isn’t flagged just for being formal.

How should I use a result?

Use it to start a talk between people, not as a verdict. If a text scores high, look at the marked sentences and the reasons. Think about the situation. If the result could matter a lot, talk to the person. Do not rely on a number alone.

How much does it cost?

ScanForAI is free to start — the scanner, accounts, history and the API. The free plan stays free. Pro and Team are paid plans with higher limits; see Pricing for the prices.

Can it clean up machine formatting?

Cleanup removes the special formatting that AI tools leave in text, like long dashes, curly quotes and hidden characters. It turns your text into plain text. It is free, and it never changes a word.

AI detection is not 100% accurate. Every result is an estimate from writing patterns and can be wrong both ways. Never use it as the sole basis for a decision about a person, and never to accuse.