Detecting-AI.com

Detecting AI provides a browser-based text detector that highlights passages with possible AI signals. Its public interface accepts pasted text and document files, alongside separate rewriting, plagiarism and fact-checking tools. Detection scores are review signals rather than proof of authorship or misconduct. Consider drafts, context and other evidence before making consequential decisions; the supplier's advertised accuracy statistics were not independently tested.

The free web plan lists fifteen thousand characters per detector check and up to three hundred detections daily without an account. Paid plans display USD14 billed monthly or USD84 charged annually, equivalent to USD7 per month. They raise detector capacity to one million characters per check. Monthly plagiarism allowances are one thousand credits; annual subscriptions provide twelve thousand credits per year. Each credit covers up to two sentences, rounded up, and each plagiarism scan accepts up to three thousand words. Unlimited detection and rewriting remain subject to ordinary human use and per-request limits. API access has separate pricing.

The actual public registration page displays empty name, email and password fields, a terms checkbox and a Google option. No account, document, scan, rewrite, OAuth authorization or payment was submitted. Paddle processes paid subscriptions, which renew until canceled; local taxes and final checkout charges require confirmation.

The privacy policy distinguishes typing from submitting a check. Extracted document text is sent for processing after submission. Anonymous checks may leave brief operational records, while signed-in history and public share links have different exposure. General model training requires separate consent according to the policy. These statements do not establish independent security certification or a fixed retention period. Avoid confidential inputs until the applicable processing and sharing settings are understood.