What are LangChain Runnables?

A Runnable is the core interface in LangChain that every composable component implements. If something is a Runnable, it can be connected with |, batched, streamed, retried, and traced — regardless of whether it's a prompt template, an LLM, a retriever, or a custom Python function.

Every Runnable exposes these standard methods:

• invoke(input) — single synchronous call, returns one output
• batch([input1, input2, ...]) — processes multiple inputs, returns list of outputs
• stream(input) — yields output chunks as they arrive (useful for token streaming)
• ainvoke() / abatch() / astream() — async equivalents of the above
• astream_events() — fine-grained async event stream (tool calls, LLM tokens, etc.)

LangChain ships several utility Runnables: RunnablePassthrough passes input unchanged (or adds extra fields), RunnableLambda wraps any Python function as a Runnable, RunnableParallel runs multiple branches concurrently and merges their outputs into a dict, and RunnableBranch routes input to different Runnables based on conditions. These combine with LCEL pipes to build arbitrarily complex workflows.