LangChain

Patterns this framework supports natively

• ★★ Tool Use — Let the LLM produce typed calls against an external toolkit instead of producing free-form text the surrounding system has to parse.
• ★★ ReAct — Interleave a single thought, a single tool call, and a single observation per step so the agent reasons over fresh evidence.
• ★★ Routing — Classify an incoming request and dispatch it to the specialist (lane / agent / model) best suited to handle it.
• ★★ Prompt Chaining — Decompose a task into a fixed sequence of LLM calls where each step's output becomes the next step's input.
• ★★ Fallback Chain — Try a primary handler; on failure or low confidence, fall through to a sequence of fallback handlers.
• ★★ Structured Output — Constrain the model's output to conform to a JSON Schema (or similar typed shape).
• ★★ Parallelization — Run independent LLM calls concurrently and combine results.
• ★★ Streaming Typed Events — Push partial results to the client as typed events as they become available, rather than waiting for the full response.