Memory Read¶

The Memory Read component provides a recall tool that retrieves information from the global memory system. When connected to an agent, the agent can look up facts by key, search by query, or list all stored memories.

Ports¶

Inputs¶

This component has no wired inputs. It operates as a LangChain tool that the agent invokes with arguments during its reasoning loop.

Outputs¶

Configuration¶

The Memory Read component accepts the following configuration in extra_config:

Usage¶

1. Add a Memory Read node from the Node Palette (Memory category)
2. Connect it to an agent node via the amber diamond memory handle
3. Optionally configure the memory type, result limit, and confidence threshold in the node's Extra Config

The agent's LLM will then have a recall tool available. The tool accepts two optional parameters:

If called with no arguments, the tool lists all stored memories up to the configured limit.

Search behavior¶

The recall tool follows this resolution order:

1. No arguments -- lists all facts from global memory
2. Key provided -- attempts exact key lookup first, then falls back to fuzzy search using the key as a query
3. Query provided -- searches across the configured memory type(s):
   • Facts -- fuzzy search with confidence filtering
   • Procedures -- finds procedures matching the goal
   • Episodes -- searches recent episode summaries
   • All -- searches across all three types

Return format¶

Results are returned as JSON arrays with the following structure:

[
  {"key": "user_name", "value": "Alice", "confidence": 0.95},
  {"key": "favorite_color", "value": "blue", "confidence": 0.8}
]

For exact key lookups, the format is a simple string: user_name = Alice.

If no results are found, a descriptive message is returned (e.g., No memory found for key: unknown_key).

Example¶

An agent with Memory Read connected can answer questions about previously stored knowledge:

User: What do you know about me?
Agent: [calls recall(query="user")]
       → [{"key": "user_name", "value": "Alice", "confidence": 0.95},
          {"key": "user_role", "value": "developer", "confidence": 0.8}]
Agent: I know your name is Alice and you're a developer.

flowchart LR
    T[Chat Trigger] --> A[Agent]
    M[AI Model] -.->|model| A
    MR[Memory Read] -.->|memory| A

    style MR fill:#f59e0b,color:white