# AI text checker: run it free, or from your agent

Highlights AI writing tells in your text with a plain-word fix for each. A style checker, not an AI detector.

## Prompt

Use looot to check this draft for AI writing tells and fix them: [paste the draft]. 1. Set up https://looot.ai/skill.md. 2. List every em dash, overused word (delve, crucial, landscape, tapestry, robust), "not just X but Y", chatbot phrase, filler phrase and forced list of three. 3. Run a chat completion endpoint to rewrite only those sentences, keeping every fact. 4. Return the list and the edited draft.

## What it needs

Up to 20,000 characters of text

## What comes back

- text with each tell highlighted
- count per kind of tell
- tells per 100 words
- a plain-word fix per tell

## Endpoints and providers

None. The tool on the web page runs in the browser and calls no endpoint.

## What it costs

Nothing. The tool on the web page runs in the visitor's browser: no account, no API call and no limit per hour. The prompt above runs the same job from your own looot account with current data, billed per call.

## Edit before you post

Catch the em dashes and stock phrases in a post or an email before anyone reads it.

## Check a rewrite

Run the AI humanizer, then paste its answer here and see what is left.

## Keep a style guide

Hold docs and release notes to plain words with a quick paste.

## Review agent output

Check what your agent wrote before it goes out, at no cost.

## Is this an AI detector?

No. It cannot tell who wrote a text, and no score here means a percent of AI. It flags style patterns people often read as machine-written. A person can write every one of them, and a model can avoid them all.

## What counts as an AI writing tell?

Eight kinds: em and en dashes, overused words such as delve, crucial, landscape and tapestry, "not just X but Y", chatbot phrases such as "I hope this helps", filler such as "in order to", lists of three, title case headings and curly quotes.

## What does tells per 100 words mean?

The number of tells found divided by the word count, times 100. It lets you compare a short and a long text. Lower is plainer. There is no pass mark.

## Why is a list of three flagged?

Models often group ideas in threes even when there are two or four. The checker flags every "A, B and C" so you can check the real number. Many will be fine.

## Is my text sent anywhere?

No. The checker runs in your browser with fixed patterns. Nothing is uploaded, and there is no account or limit.

## Does it work in other languages?

The word lists and phrases are English. Dashes and curly quotes are found in any language.

## How do I fix the tells?

Each finding has a plain fix, such as "use a comma or a period" for a dash or "use" for leverage. For a full rewrite, try the AI humanizer and check the result here.

## How does my agent use this?

Copy the prompt above. Your agent sets up looot from skill.md, lists the same tells, and rewrites only those sentences with a chat completion endpoint.

## Related tools

- /free-tools/ai-humanizer
- /free-tools/ai-text-analyzer
- /free-tools/token-counter

## Last reviewed

2026-10-14.