Product
Detection API
Send text, get a verdict back as JSON, with each sentence labelled and the detector run named.
One endpoint
POST the text and a setting to /api/v1/detect. The response carries the verdict, each sentence with its label, and the detector run that produced the reading.
Not scored under 50 words
Our detector does not score text under 50 words, and readings under 150 words are weak. Check for that before you read a verdict. The first check after a quiet spell can take longer.
Three settings
Changing the setting never turns an AI verdict into Human. It only moves the line for papers that are partly machine-written.
- "standard" flags above 6% and is the default
- "accusation-safe" flags above 15% and misses more AI-edited text than the default
- "sensitive" flags above 2% and flagged 2 of 1,928 human documents in testing
Keys and limits
Create a key in the dashboard. It is shown once and lasts 90 days. API use is not billed yet; each key is held to a ceiling on requests and words an hour, and the reference lists both numbers.
- Keys come with the school plan
No fallback
If our detector is unreachable, the response says so. It never returns a reading from a different detector.
What it reads
Our detector reads English academic and student writing. It is not built or tested for other languages or other kinds of writing.
Start with a paper you know.
Paste something you know a person wrote, such as your own writing, and see what our detector marks in it. It is free to start, with 2,000 words a day, or 1,000 papers a month with a verified school email.