How does the citation system work?

The skill tracks chunks and entities within its storage structure, enabling Claude to reference specific source documents when answering your questions.

What is the benefit of the 'hybrid' search mode?

Hybrid mode is the recommended setting as it combines local entity-level facts with global thematic analysis to provide the most comprehensive answer possible.

Does this skill require an external API key?

Yes, it requires an OpenAI API key to generate vector embeddings and process semantic queries through the LightRAG framework.

Is there a limit to the number of PDFs I can index?

The limit is primarily determined by your local storage and the token limits of your embedding model, but LightRAG is designed to handle large-scale document collections efficiently.

Can I index multiple directories of PDFs?

Yes, you can run the index command on any directory path, and the tool will build or update your knowledge graph accordingly.

Semantic PDF Research

Name: Semantic PDF Research
Author: hongsw

byhongsw

•

データサイエンスとML

Indexes PDF document collections using LightRAG to enable semantic search and context-aware research within Claude Code.

概要

Semantic PDF Research is a specialized Claude Code skill designed to bridge the gap between static document storage and interactive AI research. By leveraging LightRAG technology, it builds a comprehensive knowledge graph from your local PDF files, allowing Claude to answer complex questions with precise citations. Whether you need to extract specific technical facts or generate high-level summaries across dozens of papers, this skill provides a structured research pipeline directly in your CLI, supporting hybrid, local, and global search modes for maximum retrieval accuracy.

主な機能

Automated PDF indexing and text extraction via PyMuPDF
Context-aware answers with verifiable source citations
LightRAG integration for graph-based semantic understanding
Incremental indexing to efficiently handle new document additions
Four distinct search modes: Hybrid, Local, Global, and Naive
16 GitHub stars

ユースケース

Analyzing technical whitepapers or academic journals for implementation details
Summarizing broad themes and trends across a large library of project documentation
Performing semantic queries across local PDF collections to find facts with source references

概要

主な機能

Automated PDF indexing and text extraction via PyMuPDF
Context-aware answers with verifiable source citations
LightRAG integration for graph-based semantic understanding
Incremental indexing to efficiently handle new document additions
Four distinct search modes: Hybrid, Local, Global, and Naive
16 GitHub stars

ユースケース

Analyzing technical whitepapers or academic journals for implementation details
Summarizing broad themes and trends across a large library of project documentation
Performing semantic queries across local PDF collections to find facts with source references