How do I add a new research paper to my database?

You can add papers by running 'papi add' followed by an arXiv ID, an arXiv URL, a local PDF path, or a BibTeX file for bulk imports.

Does this skill support semantic search?

Yes, papi supports a tiered search system including literal text matching, BM25 ranked search, and full semantic RAG synthesis through the 'papi ask' command.

Can I use this to extract specific LaTeX equations?

Absolutely. You can use the command 'papi show -l eq' to instantly retrieve extracted LaTeX equations for implementation verification.

How do I ground my code implementation with citations?

Use the 'papi-ground' or 'papi-verify' sub-skills, or export context directly to your repository using 'papi export' to provide Claude with authoritative references.

When should I use the CLI instead of MCP RAG tools?

Use the CLI for fast lookups, exact term searches, and viewing known paper summaries. Escalated to MCP RAG tools only when you need semantic matching or cross-paper synthesis.

Paper Reference Assistant (papi)

Name: Paper Reference Assistant (papi)
Author: hummat

byhummat

•

データサイエンスとML

Manages a local research paper database to extract equations, summaries, and context for grounding AI-generated code in scientific literature.

The papi skill transforms Claude into a research-literate coding assistant by providing direct access to a curated database of academic papers. It allows developers to seamlessly import research from arXiv or local PDFs, perform high-speed searches for specific equations and concepts, and use RAG-powered synthesis to verify implementation accuracy. By bridging the gap between scientific papers and code, it ensures that technical implementations remain grounded in authoritative sources using a tiered search approach that prioritizes speed and cost-efficiency.

主な機能

01Context export capabilities for grounding code reviews and project documentation

02Deep integration with LEANN and PaperQA2 for formal citations and cross-paper synthesis

03Instant extraction of LaTeX equations and detailed paper summaries

04Tiered search system ranging from fast literal matching to deep semantic RAG synthesis

05Multi-source paper ingestion from arXiv IDs, URLs, local PDFs, or BibTeX files

069 GitHub stars

ユースケース

01Searching a private research library to find specific algorithmic approaches for complex problems

02Verifying that a mathematical implementation in code matches a paper's original equations

03Automatically generating implementation notes and summaries for new research during development

主な機能

01Context export capabilities for grounding code reviews and project documentation

02Deep integration with LEANN and PaperQA2 for formal citations and cross-paper synthesis

03Instant extraction of LaTeX equations and detailed paper summaries

04Tiered search system ranging from fast literal matching to deep semantic RAG synthesis

05Multi-source paper ingestion from arXiv IDs, URLs, local PDFs, or BibTeX files

069 GitHub stars

ユースケース

01Searching a private research library to find specific algorithmic approaches for complex problems

02Verifying that a mathematical implementation in code matches a paper's original equations

03Automatically generating implementation notes and summaries for new research during development