Sample Report
Every scan produces a structured report. Here is what each field means and how to use it.
| Field | What it shows | How to use it |
|---|---|---|
| AI Likelihood Score | 0–100 probability estimate | Higher score = more AI-generation signals found |
| Verdict | Likely AI / Likely Real / Unclear | Quick triage label for decision-making |
| Confidence | Low / Medium / High | How much usable evidence the scanner found |
| Evidence Frames | Timestamps of flagged frames | Review these specific moments for visual anomalies |
| Reason Codes | Machine-readable signal labels | e.g., ai_classifier_detected_ai_content, temporal_inconsistency |
| Timeline | Frame-by-frame signal distribution | See whether signals are spread across the video or concentrated |
| Compression Warning | Flag if source shows heavy re-encoding | Confidence may decrease on compressed sources |
| Final Note | Disclaimer | Results are probabilistic. Not legal or forensic proof. |
A high score with High confidence means the scanner found strong AI-generation signals across multiple frames. Review the evidence frames and reason codes for specific patterns. Even with a high score, verify the source — the video could be a real clip with heavy editing that mimics AI artifacts.
A medium score means the scanner found some suspicious signals but not enough for a strong verdict. This often happens with compressed videos, short clips, or partial AI editing. Use the evidence frames to decide whether deeper review is needed.
A low score means few AI-generation signals were found. This does not prove the video is real — it means the available evidence did not strongly suggest AI generation. High-quality AI output or heavy post-processing can produce low scores.
Ready to try it? Start a free scan or learn about how the scanner works.