I need a custom, AI-driven solution that can scan full academic manuscripts (DOCX and PDF) and flag any potential plagiarism with a high degree of accuracy. The core task is to analyse text intended for publication, recognise Chicago-style citations, and produce a clear similarity report that distinguishes properly cited passages from problematic ones. Beyond plagiarism detection, the tool must run an integrated grammar check and offer basic reference-management feedback—think automatic spotting of missing bibliographic entries or inconsistent formats. A consolidated report that merges similarity percentages, grammar issues, and reference notes into one downloadable file is essential. You are free to build in the language or framework you prefer, as long as the final deliverable includes: • A runnable application or script (desktop, web, or CLI) that accepts at least DOCX/PDF inputs • Chicago citation parsing and exclusion logic • Detailed similarity report highlighting matched sources • Built-in grammar analysis • Reference list validation Source code and a brief setup guide should accompany the finished product so I can deploy it on my own server or local machine.