Memory Registry Management¶

← API Documentation

This document describes memory registration and management for long-term knowledge storage.

Similar to workspaces, you can register directories as memories for long-term knowledge storage. Memories are persistent data stores that can be shared across agents and sessions.

register_memory()¶

Register a directory as a memory for persistent knowledge storage.

def register_memory(
    path: str,
    name: str | None = None,
    description: str = "",
    created_by: str | None = None,
    metadata: dict | None = None
) -> dict

Parameters: - path (str): Absolute path to memory directory (e.g., "/my-memory") - name (str, optional): Friendly name for the memory - description (str): Human-readable description (default: "") - created_by (str, optional): User/agent who created it (for audit) - metadata (dict, optional): Additional user-defined metadata

Returns: - dict: Memory configuration dict with keys: - path: Memory path - name: Memory name - description: Description - created_by: Creator - created_at: Creation timestamp - metadata: Additional metadata

Raises: - ValueError: If path already registered as memory

Examples:

# Register a memory for knowledge base
config = nx.register_memory(
    "/knowledge-base",
    name="kb",
    description="Shared knowledge base"
)
print(f"Registered memory: {config['path']}")

# Register with metadata
config = nx.register_memory(
    "/project-docs",
    name="docs",
    description="Project documentation",
    created_by="alice",
    metadata={"project_id": "12345", "team": "engineering"}
)

unregister_memory()¶

Unregister a memory (does NOT delete files, only removes registration).

def unregister_memory(
    path: str
) -> bool

Parameters: - path (str): Memory path to unregister

Returns: - bool: True if unregistered, False if not found

Examples:

# Unregister a memory
success = nx.unregister_memory("/my-memory")
if success:
    print("Memory unregistered")

list_memories()¶

List all registered memories.

def list_memories() -> list[dict]

Returns: - list[dict]: List of memory configuration dicts

Examples:

# List all memories
memories = nx.list_memories()
for mem in memories:
    print(f"{mem['path']}: {mem['name']}")

get_memory_info()¶

Get information about a registered memory.

def get_memory_info(
    path: str
) -> dict | None

Parameters: - path (str): Memory path

Returns: - dict | None: Memory configuration dict or None if not found

Examples:

# Get memory info
info = nx.get_memory_info("/knowledge-base")
if info:
    print(f"Memory: {info['name']}")
    print(f"Created by: {info['created_by']}")
    print(f"Description: {info['description']}")
else:
    print("Memory not found")

See Also¶

• Workspace Management - Similar registry for workspaces
• Semantic Search - Search across memories
• Configuration - YAML configuration

Next Steps¶

1. Set up semantic search for memory queries
2. Configure workspace registries