A tutorial for building autonomous agents: with LangChain and from scratch
This is a project that demonstrates the building of an autonomous agent that can search and summarize research papers from Arxiv based on user input. The project provides two examples - a highly abstracted version using LangChain for orchestration, and a detailed implementation without using LangChain.
Example 1: Abstracted Version with LangChain In this example, the agent built for searching and summarizing research papers uses LangChain for its orchestration. LangChain is a framework designed to simplify the development of conversational AI systems. The agent takes the user's research question as input, searches for relevant papers in Arxiv, and summarizes the core ideas of the paper in response.
Example 2: Detailed Implementation without LangChain In this example, we dive deep into the details and build the autonomous agent from scratch without using LangChain. We will build the reasoning step, memory, and tool-using step with pure python and OpenAI API.