EU-based · GDPR-friendly · model v1

Human or AI?
Ale-ya knows and adapts.

Paste a text. We tell you how likely it is to be AI-written, with sentence-level highlights and an honest confidence score. Flagged as AI? One click rewrites it to read human and re-checks the score. No login needed.

Live · v1
0 words · EN · auto
How it works

One model. Many signals. Honest output.

Humanizer

Flagged as AI? Make it read human.

When a text reads as AI, one click rewrites it into natural, human-sounding prose: varied rhythm, plainer wording, none of the robotic tells. Then Ale-ya re-checks the score automatically. Your meaning and language stay intact.

Use cases

Built for people who read for a living.

Honesty

Ale-ya is a probability, not a verdict.

No detector is perfect. We surface confidence honestly, never round to 100%, and show the signals behind every decision. Treat us as one source of judgement among many — and never as proof.

Pricing

Pay by detections, nothing else.

One detection = one analysis run. Unused detections do not roll over.