S
Structural AI Detector
Open Source๐ Alt to GPTZeroIdentify AI-generated web content from structure alone
๐ณ Self-Hostable๐ No Sign-upโก Traction Score: 76/100โ
66 Stars
curl -sL https://arxiv.org | head -n 25This research project trains machine learning models to detect AI-generated web content purely by analyzing structural patterns rather than textual semantics. It provides a novel approach to bypassing traditional prompt-evasion and paraphrasing tricks used to mask synthetic writing.
Evaluates raw structural markers and syntax patterns to flag synthetic content.
Bypasses standard text-flipping and semantic paraphrasing countermeasures.
Built on experimental datasets mapping distinct AI generation signatures.
Auditing large batches of newly published web articles for AI spam
Researching structural footprints left by different large language models
Enhancing search engine algorithms to downrank low-effort synthetic content
Unlike commercial SaaS detectors that rely on surface-level perplexity and burstiness, this approach targets underlying structural footprints.
AI researchers, trust and safety engineers, and search platform developers studying synthetic media detection.
Compare other trending developer tools and open-source projects in this space.