Scan your manuscript the way a publisher would. Six detection modules tell you what got flagged, which AI tool it resembles, and how to revise before you submit.
Six independent methods, not a single classifier hoping for the best.
Forensic evidence — hidden Unicode, formatting artifacts, metadata traces — revealing which AI tool touched your document. Not guesswork. Hard evidence.
Deep Scan tells you which AI model your writing resembles — ChatGPT, Claude, Gemini — so you know exactly what a publisher's tools would flag and why.
Every scan is tied to your document. Revise the flagged passages, re-scan, and watch your score drop. Track progress across drafts until your manuscript is clean.
Your documents are processed in memory and immediately discarded. ScribeGuard never stores, trains on, or retains your text.
Every flagged passage gets a severity rating, plain-English explanation, and concrete suggestion for how to revise it. Click any flag to jump straight to it.
Annotated .docx with margin comments on every flagged section and a summary report page. Your revision checklist before the editor sees it.
Upload any .docx, .pdf, .rtf, .txt, .odt, or .md file. ScribeGuard recognizes previously scanned documents automatically.
Choose Standard (our own detection algorithms) or Deep Scan (adds cloud AI analysis). Results in under 30 seconds. Your document is discarded immediately after.
Get your AI probability score, see which AI model your writing resembles, click flagged sections to jump to them, and export an annotated report as your revision checklist.
Three modules use our own algorithms. Three optional cloud modules add AI-powered analysis for the deepest inspection.
AI-isms, uniform sentence structure, repetitive transition words, passive voice overuse, and low vocabulary diversity. Adapts for literary text to minimize false positives.
Burstiness, sliding-window type-token ratio, perplexity estimation, and entropy — detecting the inhuman uniformity AI text produces.
Zero-width characters, homoglyphs, unusual whitespace, and AI tool formatting fingerprints. Logarithmic scoring avoids saturation on heavily-contaminated documents.
Evaluates writing style, coherence patterns, and voice consistency — then identifies which AI model your text most closely resembles.
AI rewrites each section, then measures how much changed. Human text transforms dramatically; AI text barely moves. The similarity ratio reveals the truth.
Evaluates cognitive load markers, self-monitoring traces, episodic memory signals, voice drift, and discourse planning patterns only human writers produce.
Most AI detectors keep your documents on their servers indefinitely. Some use uploaded text to improve their models. ScribeGuard takes a fundamentally different approach.
Your document is processed in memory and immediately discarded after analysis. We never store, log, or train on your content. Standard scans use only ScribeGuard's own detection algorithms with no third-party AI involved. Deep Scans add cloud AI analysis, but only with your explicit per-scan consent.
Start with Standard scans or go Deep for full AI-powered analysis with model identification.
30 scans/month. No third-party AI. Zero document retention.
Everything in Standard, plus 10 Deep Scan credits with full AI analysis.
Buy extra Deep Scan credits anytime from your dashboard. No subscription change needed — just top up and scan.
Shared workspace, pooled credits, batch scanning, team analytics dashboard, and admin controls.
Contact SalesPick a plan and start scanning. Cancel anytime.
By signing up you agree to our Terms of Service and Privacy Policy.
| Feature | ScribeGuard | GPTZero | Originality.ai | Copyleaks |
|---|---|---|---|---|
| AI Detection Approach | 6 layered modules | Single classifier | AI + plagiarism | AI + plagiarism |
| Zero Document Retention | ✓ processed & discarded | Stored on servers | Stored on servers | Stored on servers |
| No Third-Party AI Option | ✓ Standard tier | — | — | — |
| AI Tool Fingerprinting | Unicode & metadata forensics | — | — | — |
| AI Model Identification | Forensic + stylistic | Stylistic only | Stylistic only | — |
| Per-Document Scan History | ✓ with re-scan tracking | Via Chrome extension | Scan log | Scan reports |
| Export Annotated .docx | ✓ with margin comments | — | — | PDF export |
| Plagiarism Checking | — | ✓ Premium | ✓ | ✓ |
| LMS Integrations | — | Canvas, Moodle, etc. | — | Canvas, Moodle, Blackboard, etc. |
| Starting Price | $15/mo (30 scans) | $14.99/mo (150K words) | $14.95/mo (pay-per-check) | $7.99/mo (1,200 credits) |
Pricing and features based on publicly available information as of March 2026.