You can now cache the results of tasks in LangGraph based on node input, helping you avoid redundant computation and speed up execution.

How it works:

• Define a cache at compile time or entrypoint

• Set cache policies per node:

  • key_func to control cache key generation (defaults to input hash)

  • ttl to control cache expiration (or leave unset for persistent cache)

See the docs: https://langchain-ai.github.io/langgraph/concepts/low_level/#node-caching