Under the hood

Built to be read, not just trusted

Per-sentence evidence

Every flagged sentence comes with the specific analyzer and threshold that flagged it. There is no confidence score handed down from an opaque model: you see the readability variance, the vocabulary metric, the burstiness figure, or the stylometric deviation that triggered the flag, and you can decide for yourself whether it holds up.

Genre-tuned detection for every content type

Choose a genre preset for your content type, academic papers, marketing copy, or general text, then fine-tune thresholds like analyzer weights, sensitivity bands, phrase density, and correction tolerance.

Balanced

Default sensitivity. Good all-round detection with a moderate false-positive rate.

Strict

Aggressive detection. Catches more AI text but may flag polished human writing.

Lenient

Relaxed thresholds. Fewer false positives, lets borderline text through.

Academic

Tuned for scholarly text. Less sensitive to formal, polished writing style.

Marketing

Tuned for product pages and commercial copy. Less sensitive to repeated domain vocabulary.

Genre presets and threshold tuning available on Professional and Enterprise plans.

Branded PDF detection reports

Export a full detection report as a PDF: composite score, per-analyzer breakdown, flagged sentences with evidence, and, if correction ran, a before-and-after summary. Built for sharing with an editor, a client, or a review committee that needs a paper trail, not just a percentage.

REST API

Submit text and get back the same structured detection result the web app uses: composite score, per-analyzer scores, and evidence, as JSON. Enterprise plans get full API access for integrating detection into an existing editorial or moderation pipeline.

Chrome extension

Right-click any selected text on any webpage to analyze it instantly, no copy-paste, no tab-switching. Detection lives where you already read.

Install extension