id: langgraph_persistence_pinned
framework: langgraph
mode: white_box
status: implemented_store_harness
description: LangGraph persistence target using InMemoryStore plus an explicit benchmark write/retrieval policy.
source_url: https://docs.langchain.com/oss/python/langgraph/persistence
source:
  type: python_package
  package: langgraph
  version: 1.2.7
runtime:
  type: local_python
variants:
  - langgraph_store_enabled
model:
  provider: none
  model: not_used
  temperature: 0
adapter:
  module: memorybench.adapters.langgraph_persistence
  class: LangGraphPersistenceAdapter
capabilities:
  inspect_memory: true
  delete_memory: true
  retrieval_log: true
  multi_user: true
  multi_tenant: false
notes:
  - LangGraph provides persistence primitives; this result measures the store with MemoryStackBench's minimal adapter policy, not a built-in semantic memory agent.