Framework Selection

Framework selection depends on project complexity: LangChain for single-purpose agents, LangGraph for complex control flow orchestration, and Deep Agents for feature-rich agentic environments. Use decision criteria matching your specific feature requirements to determine the right AI architecture layer.

The layered architecture illustrates how these frameworks build upon each other. LangChain provides the foundational components for single-purpose agents, while LangGraph builds on top to enable complex control flow orchestration for multi-step AI workflows.

You need LangGraph when your AI assistant requires complex control flow orchestration beyond simple single-purpose agents. If your project involves multi-step workflows with intricate state management, LangGraph provides the necessary architecture to handle that complexity.

Evaluate Deep Agents when your project demands a feature-rich agentic environment rather than simple task execution. If your AI architecture requires extensive built-in capabilities and advanced agent interactions, Deep Agents provides the comprehensive environment needed for those feature requirements.

Yes, this framework selection guide is designed for developers starting new AI assistant projects. It provides clear decision trees and framework profiles to help you determine the optimal architecture from the outset, ensuring efficient development whether you need LangChain, LangGraph, or Deep Agents.