new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

LangChain · Agents · all subjects

state & checkpointing

2 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Agent invocation with checkpointer for memory

Agents support stateful execution with a checkpointer parameter. In Python: create_agent(..., checkpointer=checkpointer). When invoking, pass config={'configurable': {'thread_id': 'unique-id'}} to enable conversation state. In TypeScript: createAgent({..., checkpointer}) and invoke(..., {configurable: {thread_id: 'unique-id'}}). The MemorySaver (JS) or InMemorySaver (Python) can be used for in-memory checkpointing. This allows agents to remember previous conversations within the same thread.

InMemorySaver and MemorySaver for agent memory

InMemorySaver (Python from langgraph.checkpoint.memory) and MemorySaver (TypeScript from @langchain/langgraph) provide in-memory checkpointing for agent state. They store conversation history and context during agent execution. Pass checkpointer=InMemorySaver() (Python) or checkpointer=new MemorySaver() (JS) to create_agent/createAgent. When invoking agents, use config={'configurable': {'thread_id': 'id'}} to persist state to a specific thread. For production, use persistent checkpointers that save to a database.

Give your agent this brain