Detecting-AI.com vs GPTKit

Detecting-AI.com analyzes writing for signals associated with AI-generated text and presents a document score with highlighted passages. The web detector accepts pasted text plus TXT, PDF, and DOCX files; scanned PDFs requiring OCR are not supported. It is designed for text linked to GPT, Claude, and Gemini, including edited or AI-assisted drafts. The service preserves document structure, scores eligible sentences separately, and combines those scores into a document signal. The maker reports 99% detection accuracy and a false-positive rate under 1% on its described datasets, while noting that short, translated, heavily edited, or formulaic writing can be harder to classify and that a score cannot establish authorship or misconduct. Paid plans add unlimited AI detection, an AI Humanizer, Fact Checker, and plagiarism credits. REST APIs cover detection, humanization, and plagiarism checks, with a remote MCP server for compatible assistants. The free plan requires no account and allows 15,000 characters per check. Paid pricing starts at $7 a month billed annually.

GPTKit analyzes submitted text and classifies it as human-written or ChatGPT-generated. It is available through a web interface and an API, with English currently supported. The maker says the detector combines six AI-based methods, was trained on 1 million datasets, and produces authenticity and reality reports. GPTKit reports approximately 93% accuracy based on testing 100K+ human and ChatGPT responses. The first 2048 characters of every request are free, including for guests; the Free Plan allows up to five requests per minute. Unlimited (Personal) costs $5 a month and supports two team members. Unlimited (Team) costs $10 a month, supports eight members, and includes API access. Unlimited (Institution) costs $20 a month, supports 20 members, and includes API access. Team credits can be shared. GPTKit says submitted text is temporarily stored and removed after detection; its policy says absolute security cannot be guaranteed.