ArXini
ArXini is an AI agent that bridges the gap between academic research and practical implementation.
Core trick
Links an arXiv paper to its GitHub repository, indexes the codebase into a vector store, and answers questions over both with Gemini.
Description
ArXini is an AI agent that bridges the gap between academic research and practical implementation. It uses Google Gemini to analyze any ArXiv paper and its GitHub repository, allowing users to interactively explore how theoretical concepts are translated into code. ArXini transforms a complex research task into a simple, on-demand service and allows you to get onboarded / start working on existing implementations much faster.
README excerpt
# Repo Rover (ArXini) **Deep Research, Decoded** — An AI-powered application that bridges academic research papers to their code implementations, enabling researchers and developers to understand paper-to-code connections through intelligent Q&A.  ## What It Does Repo Rover takes an academic paper (from ArXiv), finds the GitHub repository that implements it, indexes the codebase using semantic search, and lets you ask natural language questions about how the paper's concepts are implemented in code. **Example workflow:** 1. Search for "Attention Is All You Need" 2. The system finds the paper on ArXiv and discovers its GitHub implementation 3. The codebase is indexed into a vector database 4. Ask: *"How is the multi-head attention mechanism implemented?"* 5. Get an AI-synthesized answer linking theory from the paper to actual code snippets ## Features - **Paper Search** — Search by title, ArXiv ID, or URL with Gemini AI-powered discovery - **Automatic Repo Discovery** — Multimodal PDF analysis extracts GitHub links from papers - **Semantic Code Indexing** — ChromaDB with Jina embeddings for intelligent code retrieval - **RAG-Powered Q&A** — Ask questions and get answers that connect paper theory to code - **Concept Mapping** — AI-generated maps linking paper concepts to code implementations - **PDF Viewer** — Read the paper side-by-side with the chat interface - **Voice Input** — Record spoken questions with Gemini audio transcription - **Sho