emend.it is developing a language-based writing support system designed to refine and rework text in ways that reflect natural human writing practices rather than robotic ones. This predates the widespread use of generative artificial intelligence. The system concentrates on sentence construction, lexical precision, tonal balance, rhetorical restraint, and stylistic variation characteristic of human-authored academic and professional writing. The purpose is to make sure that the content generated by the user feels natural to the way the user writes and the tone of the user’s preference. It is meant to enhance clarity, coherence, and readability while preserving the author's original meaning, argumentation, and intellectual ownership.

emend.it is intended for academic, professional, and institutional writing contexts where authors seek to improve flow, decrease the patterns of the generative AI and reduce formulaic or mechanistic phrasing, as well as align the work of the user with established human writing conventions. The system does not generate content independently, nor fabricate references, neither does it submit work on behalf of users, but it functions strictly as a revision and paraphrasing aid that supports ethical authorship and responsible writing practices and less robotic writing.

Contributors who open this page are willing to submit pre-AI written text and materials to support the training of the model and may choose whether they should be acknowledged on a public contributor's page and will be invited to join a priority waitlist to access and evaluate emend.it prior to public release to test the model. Contributors are to read the terms and conditions below and agree to it before submitting their original content for the model training.

emend.it will be of support to both academic and professional writing by redefining the language and style of AI-written text using patterns observed in human-authored text which were produced before the widespread adoption of generative artificial intelligence. It will function as a writing support rather than an authorship replacement.

By agreeing to submit material to this project, you confirm that the content you uploaded is your original work or that you hold the legal right to share that document for research purposes. Documents submitted are solely for the purpose of training, evaluation and refinement of the language model and its associated retrieval process that analyses aggregate patterns of the human writing style. The system will not store or reproduce individual submission verbatim in outputs, nor will the submissions be used for marketing, resale or public redistribution. Where feasible, personal identifiers are separated from textual data or replaced with a pseudonym.

The generated output of the model, once training is complete, remains the responsibility of the user, who is expected to review and edit to ensure compliance with academic integrity policies and professional standards. All submitted data will be stored securely and accessed only by the research team. Submitted content will not be shared with third parties. Contributors may request deletion of their submission at any point prior to public deployment of the system.

Participation is voluntary and uncompensated, but contributors may opt to be acknowledged publicly and invited to participate in early evaluation phases of the system. Submission of content constitutes informed consent to these terms.