AI text detection
Examines vocabulary repetition, recurring phrases, sentence-length distribution, punctuation frequency, and unusual characters. More complete text usually produces steadier statistical signals.
Open text detectorConsent before upload · Inspectable evidence
Free AI content detector Text · Image · Video
No login or sign-up required. Available services are selected automatically, with consent before image or video uploads. Text currently provides local supporting clues; uncalibrated results remain inconclusive.
Content screening and verification
Results separate heuristic screening signals, file facts, and evidence limitations.
Text content
GenDetect checks repetition, sentence length, punctuation, and character structure.
Image content
Automatically checks generator metadata, provenance credentials and model signals, with consent before server analysis.
本地启发式不会上传图片。
Video content
Server analysis supports videos up to 120 seconds and 100MB, inspecting at most six sampled frames after upload consent. Audio and temporal continuity are not checked.
Detection methods
Each content type needs a different analysis. GenDetect separates screening signals, file facts, and limitations.
Examines vocabulary repetition, recurring phrases, sentence-length distribution, punctuation frequency, and unusual characters. More complete text usually produces steadier statistical signals.
Open text detectorChecks original-file metadata, provenance credentials and model signals. Unverified claims and uncalibrated scores remain supporting clues.
Open image detectorServer analysis checks at most six frames of videos up to 120 seconds. It does not inspect audio or temporal continuity, and a negative result does not establish human origin.
Open video detectorHow to read the score
This demo locally aggregates heuristics for text, images, and sampled video frames; images also check known generator metadata. The index is not trained or probability-calibrated and cannot prove origin.
Understand and verify
Understand the signals and false-positive limits first, then verify the original file, provenance, and editing process. These resources separate methodology, accuracy, and human review.
See how text and image signals form screening indexes, which video file facts are read, and what browser heuristics cannot determine.
Read methodology and limitationsUnderstand false positives, false negatives, thresholds, and sample differences without mistaking an uncalibrated screening score for a probability of origin.
Review accuracy and false positivesCheck original material, content credentials, publishing sources, version history, and contextual facts with separate workflows for text, images, and video.
Read the human verification guideFrequently asked questions
Understanding the limits of detection scores helps reduce false positives for human and edited content.
AI content detection provides supporting clues only. Models, rewriting, compression, and input length can all affect the result, so a score alone cannot establish the source.
Longer text contains more samples of sentence structure, repetition, and punctuation, making statistical signals steadier than they are for short snippets.
Images are checked for metadata, provenance and model signals. Video analysis checks at most six sampled frames, without audio or temporal analysis. Uncalibrated model scores cannot independently establish origin.
No. Verify the author, publication date, original file, content credentials, and publishing source before reaching a conclusion.
No login or sign-up is required. Detection follows enabled service capabilities. Image and video server analysis requires upload consent; text stays in the browser when specialist text detection is unavailable. Completed uploads are deleted, interrupted uploads expire within one day, and any external providers are disclosed before upload.