LLM Agents Team Up: The Future of Automated Collaboration
What’s better than one LLM agent? Multiple working together! Well… maybe. Multi agent collaboration is an LLM agent based framework, where multiple LLM Agents with their own respective tools work together to deliver on a given task, without the need for a supervisor. In this, we break down the framework of how this team collaborates with each other, using three distinct agent workers: A weather lookup agent, an events lookup agent, and a web search agent- all working to give you the perfect weekend plans and overview!
Code Available Here: https://github.com/ALucek/multiagent-...
Resources:
LangChain Video Tutorial on Multi Agent Frameworks: • LangGraph: Multi-Agent Workflows
LangChain Github Repo for Multi Agent Collaboration: https://github.com/langchain-ai/langg...
LangChain Blog Post about Multi Agent Workflows: https://blog.langchain.dev/langgraph-...
OpenWeatherMap API: https://openweathermap.org/api
TicketMaster API: https://developer.ticketmaster.com/pr...
Chapters:
00:00 - Intro
00:41 - Diagram & Flow Overview
03:36 - Code Walkthrough Starts
04:00 - Defining Tools
06:00 - Defining Custom Tools: Weather & Events
08:05 - Defining Custom Tools: Schema & Tool Functions
09:39 - Helper Functions: Agent Creation & Node Creation
12:09 - LangGraph State Setup
13:06 - Creating Three Agents
14:14 - Creating Agent Node & Tool Use Node
14:58 - Routing Logic Function
15:44 - Compiling LangGraph Graph
18:35 - First Demonstration!
20:59 - Pros & Cons
22:40 - Second Demonstration!
24:16 - Outro