@ltryee

ocr-locally extracts text from images and pdfs on macos using native vision and pdfkit frameworks. supports multi-language recognition, confidence scoring, and json output modes without external dependencies.

xinwenlianbo-translator decodes cctv official rhetoric into accessible chinese for educated audiences. fetches daily transcripts, applies jargon mapping, and produces structured summaries with policy signal analysis.

deep-reader applies systematic reading methodology to analyze books and web articles across four depth levels, from basic recognition through comparative thematic analysis, with output formats keyed to user intent.

obsidian-assistant targets knowledge base management by recording user habits, understanding vault structure, and extracting reusable templates and dataview queries. coverage is narrowly focused on obsidian workflows without cross-vendor applicability.