framework-selection

Framework selection depends on task complexity. Use LangChain alone for simpler chains, LangGraph with LangChain for stateful multi-step workflows, or Deep Agents for complex autonomous tasks, potentially mixing layers for specific needs.

You need LangGraph when your workflow requires complex state management or cyclical multi-step agent logic. LangChain alone handles linear chains, while LangGraph adds stateful orchestration for dynamic task routing.

The best way to start is using a quick-start protocol to map project requirements to framework profiles. This ensures you correctly align the development approach and setup whether using Deep Agents, LangChain, or a hybrid stack.

Yes, you can mix layers. The decision guide provides mixing-layer guidance to map project needs to the appropriate stack, ensuring correct setup and future extensibility when combining LangChain and LangGraph.

Selecting the framework layer at startup reduces misalignment and rework. It ensures correct setup and future extensibility by aligning the development approach with the chosen LangChain, LangGraph, or Deep Agents architecture.

Yes, framework selection works for autonomous tasks by recommending Deep Agents or hybrid approaches. It provides decision guidance to map high task complexity to the appropriate framework layer for correct setup.