dsh-paperlens

When writing a thesis or surveying the literature, the real time sink is rarely the reading itself. It is bouncing between databases, hand-formatting citations, and juggling browser bookmarks with scattered spreadsheets to track what has been read. dsh-paperlens pulls all of those fragmented steps into a single deepseek-harness conversation: one natural-language prompt kicks off searching, deduplication, citation formatting, bulk DOI import, saving, and export without ever leaving the chat window.

The plugin ships six tools: cross-source paper search across OpenAlex and CrossRef with automatic deduplication and a metadata-completeness score; single-paper detail retrieval; one-click citation in five styles (APA 7th, MLA 9th, GB/T 7714, Chicago 17th, Vancouver); batch DOI import with metadata enrichment; a local SQLite library with tags and reading status (unread, reading, read); and export to RIS, BibTeX, or CSV. All data flows through open academic APIs with no API key required, and the library lives on your machine, so nothing leaves your hardware.

It fits three audiences well: Chinese-thesis and journal authors who need GB/T 7714 formatting; graduate students and researchers who want literature management embedded in their AI conversation workflow to cut tool-switching; and privacy-conscious users who prefer a local-first stack over uploading their paper library to a third-party SaaS. Deep AI analysis such as comparative summaries is handled natively by the dsh session model; the plugin itself only supplies structured data and does not bring its own AI gateway into the loop.