ScribeLens checks essays, reports, and other long drafts for AI writing patterns, including text written with Claude. It marks each sentence and names the reason for every flag, so you know where to look in a long document.
Many people use Claude for longer, analytical work: essays, reports, technical explanations, and arguments that cover a topic in depth.
This kind of writing is thorough and evenly organised: a claim, then supporting detail, clear signposting between sections, and counterpoints in turn. Good human writing has these qualities too, so ScribeLens shows the evidence for each sentence and lets you check it.
Detector v4.0 is the live version, and every report names the version that scored it. In our own testing, v4.0 is markedly better at detecting AI patterns in text written with current assistants such as Claude.
The result tells you which sentences show AI-like patterns and why. It does not name the tool that wrote the text, because assistants share these patterns. We retrain the detector as writing tools change.
In a long document, the sentence view shows you where to look. Flags gathered in one section, such as a summary or a run of transitions, tell a different story from flags spread evenly across a whole report.
Educators reading polished submissions, editors screening analytical pieces, and teams reviewing drafts before publishing can go straight to the flagged sentences. The reason keyword on each one names the pattern, and reference lists and citations are set aside automatically.
No. A result is evidence, not proof. ScribeLens shows which sentences carry AI-like patterns, the reason for each flag, and a graded verdict. You weigh that together with what you know about the writer.
Style varies with the model, the prompt, and how much a person edits afterwards. ScribeLens reads for the patterns that assistants share: even sentence rhythm, predictable word choice, formulaic structure. That is why one check covers text written with Claude and with other assistants.
Yes. We retrain the detector as writing tools change. Detector v4.0 was retrained in October 2026, and in our own testing it is markedly better at detecting text written with current assistants such as Claude.
No. A low score means the text did not show strong AI-like patterns in this scan. AI text that has been heavily rewritten or paraphrased is harder to detect, so weigh a low score together with drafts and what you know about the writer.
Yes. A free account scans a whole report or thesis chapter in one go, within a 35,000-word monthly allowance. Without an account a scan covers up to 2,000 words.
Paste an essay, report, or structured draft and see the sentence-level evidence behind the result. Without an account a scan covers 2,000 words; a free account scans the whole document.