Back to Workshop
AI Chat Agent Platform

AI Chat Agent Platform

Spec Sheet

ReactPythonLangChainWebSocket

AI Chat Agent Platform

The AI Chat Agent Platform was created to solve the challenge of executing complex multi-step reasoning tasks. It implements a supervisor-worker hierarchy using LangChain, allowing a coordinator agent to delegate sub-tasks (like searching the web or computing equations) to specialized tool-enabled agents. It handles real-time token streaming and manages chat history across multiple sessions dynamically. By separating concerns, it prevents context bloat in single LLM calls, ensuring high accuracy and quick responses.

Key Features:

  • Dynamic agent orchestration using custom supervisors
  • Persistent state management with Redis-backed message history
  • Streaming responses with token-by-token WebSocket propagation
  • Integrates web search, calculator, and custom code executing sandbox tools