Detecting-AI.com
detecting-ai.com
- Free plan
- Web
- API
What Detecting-AI.com is
Detecting-AI.com analyzes text for signals associated with AI writing and displays a document score with AI-like passages highlighted. The web detector accepts pasted text and TXT, PDF, or DOCX files; scanned PDFs that require OCR are not supported yet. It is designed to review text associated with GPT, Claude, and Gemini, including human-edited and AI-assisted drafts. The service scores eligible sentences separately and averages them into a document signal. The maker reports 99% detection accuracy and a false positive rate under 1%, while noting that short, translated, heavily edited, or formulaic writing can be harder to classify. A score cannot establish authorship or misconduct. The free plan costs nothing and allows up to 300 detections per day, with a 15,000-character limit per check. Yearly costs $7 a month, billed annually at $84 once per year; Monthly costs $14 a month, billed monthly. API access is available, including REST APIs for detection, humanization, and plagiarism checks.
Who it suits
It may suit educators, editors, publishers, or individuals reviewing text with care. It also offers API-based deployments and integrations for organizations.
Good to know, and what to watch
- Accepts pasted text and TXT, PDF, or DOCX files
- Highlights passages and provides a document score
- Free plan allows up to 300 detections per day
- REST APIs cover detection, humanization, and plagiarism checks
- Scanned PDFs requiring OCR are not supported yet
- Short, translated, heavily edited, or formulaic writing can be harder to classify
- A score cannot establish authorship or misconduct
- Free checks are limited to 15,000 characters each
Plans
Every plan and priceFree
Monthly
Yearly
Free plan
The verdict
Detecting-AI.com provides text scoring, highlighted passages, and API access, with stated limits and classification caveats. Its results should not be treated as proof of authorship or misconduct.
What the maker publishes
- Accuracy claim
- The maker reports 99% detection accuracy and a false positive rate under 1%, measured on diverse datasets it describes on its site. source
- API
- The developer documentation provides REST APIs for detection, humanization, and plagiarism checks, plus a remote MCP server for compatible assistants. source
- Audience
- The maker names education, editorial review, publishing, and personal writing as use cases for reviewing text with care. source
- Data use
- The maker says it does not sell submitted text or use it to train a general-purpose AI model unless the user separately and expressly agrees. source
- File input
- The web detector accepts pasted text and TXT, PDF, or DOCX files; scanned PDFs requiring OCR are not supported yet. source
- Headquarters
- International House, 10 Churchill Way, Cardiff, United Kingdom, CF10 2HE source
- Integrations
- The maker describes API-based Moodle deployments scoped with each institution and a Zapier app with detector, humanizer, and plagiarism actions. source
- Limits
- The site says short, translated, heavily edited, or formulaic writing can be harder to classify, and a score cannot establish authorship or misconduct. source
- Method
- The website preserves document structure, scores eligible sentences independently, and averages their scores into a document signal. source
- Model coverage
- The detector is designed to review text associated with GPT, Claude, and Gemini, including human-edited and AI-assisted drafts. source
- Privacy
- The privacy policy says text is not sent while typing, supported files are read in the browser, and extracted text is sent only after a check starts. source
- Purpose
- The detector analyzes text for AI writing signals and displays a document score with AI-like passages highlighted in the full text. source
- Security
- The privacy policy lists encrypted connections, restricted server-side credentials, secure HTTP-only session cookies, request-size and rate limits, response validation, and bot protection. source
- Support
- The contact page directs product, account, billing, API, integration, and partnership inquiries to [email protected]. source
Checked against the maker's pages on 7 Oct 2026.