AI text detector

Detection sends the submitted text to NovelThread’s servers. The LLM method forwards it to the configured model provider; the HF classifier, when enabled, runs on our server with Chinese and English models. Available methods are shown in the tool. Neither method proves authorship; short, translated or edited passages can produce false positives and false negatives. Six interface languages are not evidence of detection accuracy in six languages. Use the result as a prompt for discussion, never as sole evidence against an author.