How it works
A straightforward process designed to support your judgment, not replace it.
Paste or upload text
Copy and paste the text you want to check, or enter a URL. Works with essays, coursework, articles, or any written content. No file size limit for text.
AI risk pattern analysis
Multiple detection models analyse the text simultaneously. Each model uses different techniques to identify patterns commonly associated with AI-generated writing. This multi-model approach reduces false positives and gives more reliable results.
Review flagged sections
See which specific passages triggered detection. Rather than just an overall score, you can focus on particular sections that may warrant further review. Each detector's results are shown transparently.
Generate a clear report
Download a report showing the overall risk score, individual detector results, and flagged passages. Teachers can use this as evidence to support conversations. Students can use it to identify areas to revise.
Use the output as one signal
Results are probabilistic, not definitive. They should inform your judgment, not replace it. For teachers: combine with writing history, in-class work, and conversations. For students: use to understand and revise.
Important: what this tool is and is not
This tool is:
- ✓A screening signal to support your judgment
- ✓A way to identify passages that may need further review
- ✓Evidence to support fair conversations about academic integrity
- ✓Privacy-first: student data is never stored
This tool is not:
- ✗Definitive proof of AI use
- ✗A replacement for professional judgment
- ✗100% accurate (no AI detector is)
- ✗Grounds for accusation on its own
Our approach to detection
Multi-model detection
We combine results from multiple independent AI detection models. Each uses different techniques and has different strengths. When all models agree, confidence is high. When they disagree, we flag the uncertainty clearly.
Transparent scoring
We show you each detector's individual result alongside the combined score. No black boxes. You can see exactly what each model thinks and make your own assessment.
Privacy by design
Content is processed in real-time and immediately discarded. We do not store, train on, or share the text you submit. Student work remains private.
Honest about limitations
No AI detector is perfect. We are upfront about this. False positives happen. Writing styles vary. That is why we position this as a screening tool, not a verdict.