langmem0 Modules

ChatOpenAI

pydantic model langmem0.ChatOpenAI[source]

Chat model that uses the Mem0-based OpenAI API.

Config:

  • extra: str = ignore

  • protected_namespaces: tuple = ()

  • arbitrary_types_allowed: bool = True

  • populate_by_name: bool = True

  • validate_by_alias: bool = True

  • validate_by_name: bool = True

Validators:

  • with_mem0 » all fields

field mem0: dict[str, Any] [Required]

The Mem0 configuration to use.

field user_id: str | None = None

The user ID to use for the Mem0 API.

Mem0Middleware

class langmem0.Mem0Middleware(config)[source]

Bases: AgentMiddleware

Middleware for integrating Mem0 memory with LangChain agents.

This middleware automatically stores conversations and retrieves relevant memories during model calls to provide personalized responses.

__init__(config)[source]

Initialize the Mem0 middleware.

Parameters:

config (dict[str, Any]) – Mem0 configuration dictionary.

async aafter_agent(state, runtime)[source]

Async handler called after agent execution.

Parameters:

  • state (AgentState) – The agent state.

  • runtime (Runtime) – The runtime context.

Returns:

Returns None if user ID is not found.

Return type:

None

after_agent(state, runtime)[source]

Handler called after agent execution.

Parameters:

  • state (AgentState) – The agent state.

  • runtime (Runtime) – The runtime context.

Returns:

Returns None if user ID is not found.

Return type:

None

async awrap_model_call(request, handler)[source]

Async wrapper for model calls with memory context.

Parameters:

  • request (ModelRequest) – The model request.

  • handler (Callable) – The next handler in the chain.

Returns:

The result of the model call.

Return type:

ModelCallResult

wrap_model_call(request, handler)[source]

Wrapper for model calls with memory context.

Parameters:

  • request (ModelRequest) – The model request.

  • handler (Callable) – The next handler in the chain.

Returns:

The result of the model call.

Return type:

ModelCallResult