FareedKhan-dev/langgraph-long-memory ? reverse-engineered prompt

Build me a notebook that walks through a working example of long term memory for an agent with LangGraph.

I want to see how short term conversation memory and cross session memory fit together, using the in memory store for testing. Please set up a simple agent that can remember user preferences or facts from earlier chats, store them, and pull them back later in a new thread. Include the basic graph flow, the prompts, the memory helper functions, and a simple human in the loop step where the user can accept or edit what the agent remembers.

Make it easy to run in one notebook from top to bottom, with clear sections and a couple of test examples that show the memory actually working. If you need to check the current LangGraph docs to match the latest way of doing this, go ahead and look them up online.