Chat With Arxiv

This builds a conversational research assistant that searches ArXiv, retrieves papers, and answers questions about them with proper citations. It combines paper discovery through the ArXiv API, PDF processing to extract content, and RAG-based QA to pull relevant sections when responding. The multi-turn conversation system classifies whether you're asking about a single paper, wanting synthesis across multiple papers, or tracking research trends. The architecture is solid for literature review tools or academic chatbots. One thing to watch is PDF extraction quality, it varies wildly depending on how papers are formatted, so you'll want robust error handling and maybe fallbacks to abstract-only answers when full text parsing fails.

Install to Claude Code

npx -y skills add qodex-ai/ai-agent-skills --skill chat-with-arxiv --agent claude-code