AI_paper

AI_paper is a Windows desktop application for preparing and revising academic documents. It is intended for users who need help with academic writing, document refinement, AI-generated text detection reduction, plagiarism reduction, and proofreading. The application combines these tasks in one local writing workflow rather than requiring separate tools for each stage.

For academic writing, AI_paper can generate outlines, draft chapters step by step, and produce abstracts, keywords, and references. It supports paragraph optimization and rewriting, citation formats including GB/T 7714, APA, and MLA, and a rich text editor with formatting options such as bold, italics, underlining, color, and indentation. Its AI detection reduction tools scan text by segment, provide intelligent rewriting at Light, Deep Rewrite, or Academic Style intensity, and can import DOCX or PDF detection reports with paragraph-level risk labels and color-coded results. Plagiarism reduction includes Light, Medium, and Deep rewriting levels, local duplicate-fragment detection, AI rewriting, and import of plagiarism reports for locating duplicate content.

Academic polishing can be applied according to text type, including abstracts, introductions, body sections, conclusions, and outlines. It provides Grammar Correction, Vocabulary Enhancement, Logic Optimization, and Comprehensive Polishing modes, along with four execution methods. Intelligent proofreading covers Basic Text, Grammar & Syntax, Academic Formatting, Citation Standards, Data Expression, Logical Rigor, Compliance Risks, and AI Writing Traces, with Notice, Warning, and Error severity levels and one-click automatic fixes. The application also saves up to 200 operation records, supports filtering, difference comparison, and one-click rollback, and provides bidirectional conversion between text and LaTeX.

AI_paper can import DOCX files directly for editing and export the completed work, fitting the document-focused workflow of a Windows academic writing application. Data is stored locally, and API keys are encrypted, while registration and internet login are not required for use. MirrorNest links directly to the official release instead of hosting a mirrored copy, and the software license is currently unknown.