Core backend logic for a multi-agent literature-review assistant powered by the AutoGen AgentChat stack (β©Ύ v0.4).
This module coordinates a two-agent team to automatically search, filter, and summarize research papers from arXiv into a concise Markdown-formatted literature review.
- Automated Paper Discovery β Queries arXiv via the official
arxivPython API. - Agentic Workflow β Two collaborating LLM agents:
- π΅οΈ Search Agent β Crafts queries, calls arXiv, and filters the results.
- βοΈ Summarizer β Writes a structured Markdown review.
- Portable β Self-contained design, easily embedded in:
- Streamlit dashboards
- Streaming Output β Yields conversation messages incrementally for real-time display.
- Clone the repository
- Install dependencies: