This directory explains how to integrate MemMachine memory operations into n8n workflows. It provides guidance on using MemMachine’s nodes within n8n.

MemMachine provides custom nodes that integrate with n8n workflows, enabling AI agents to have persistent memory capabilities. This allows agents to retain past interactions, user preferences, and context across multiple sessions.

To add MemMachine’s community nodes from npm to an existing n8n instance:

1. Go to Settings > Community Nodes.

2. Select Install a community node.

   

3. Enter @memmachine/n8n-nodes-memmachine in npm Package Name.

4. Agree to the risks of using community nodes: select I understand the risks of installing unverified code from a public source.

5. Select Install.

   

To run a self-hosted n8n instance with MemMachine’s nodes pre-installed:

1. Go to the MemMachine n8n integration repository and clone it.
2. Start n8n with MemMachine’s nodes:

# Start n8n and Jaeger
docker-compose up -d

# View logs
docker-compose logs -f

3. Access services:

• n8n UI: http://localhost:5678
• Jaeger UI: http://localhost:16686

4. Stop n8n instance:

# Stop services
docker-compose down

5. Refer to the README.md for detailed usage instructions.

MemMachine offers two n8n nodes: the Main Node (MemMachine Manager) and the AI Memory Node (MemMachine Memory).

The Main Node provides two actions: Store a message and Enrich with context.

• Store a message

  Save conversational messages to MemMachine.

  Parameters

  Parameter	Type	Default	Description
  orgId	string	-	Required. Unique organization identifier.
  projectId	string	-	Required. Unique project identifier.
  types	multiOptions	['episodic', 'semantic']	Required. Memory types to use.
  producer	string	-	Required. Who created this message.
  producedFor	string	-	Required. Intended recipient of the message.
  episodeContent	string	-	Required. The message content to store.
  episodeType	options	''	Type of episode. Possible value: 'message' (leave empty for default).
  metadata	json	{}	Additional metadata to associate with the message.
  sessionId	string	-	Required. Unique session identifier.
  groupId	string	default	Required. Unique group identifier.
  agentId	string	-	Required. Unique AI agent identifier.
  userId	string	user_v1	Required. Unique user identifier.

  Output

  {
    "results": [
      {
        "uid": "New memory ID"
      }
    ]
  }

• Enrich with context

  Retrieve historical context from MemMachine and enrich AI agent responses.

  Parameters

  Parameter	Type	Default	Description
  orgId	string	-	Required. Unique organization identifier.
  projectId	string	-	Required. Unique project identifier.
  types	multiOptions	['episodic', 'semantic']	Required. Memory types to use.
  query	string	-	Required. Natural language query used to search for relevant memories.
  limit	number	50	Maximum number of results to return.
  filter	string	-	Filter expression to refine memory search results.
  expandContext	number	0	Number of extra episodes to include for context.
  scoreThreshold	number	-	Minimum relevance score required to include a memory.
  sessionId	string	-	Required. Unique session identifier.
  groupId	string	default	Required. Unique group identifier.
  agentId	string	-	Required. Unique AI agent identifier.
  userId	string	user_v1	Required. Unique user identifier.
  enableTemplate	boolean	true	Whether to render a formatted context template using categorized memories.
  contextTemplate	string	default template	Markdown template for formatting the memory context.

  default template refers to the system’s built-in context template. Users may override it with a custom template.

  Output

  {
    "episodic_memory": "All episodic memories, including short-term and long-term",
    "shortTermMemory": "Short-term episodic memory results",
    "longTermMemory": "Long-term episodic memory results",
    "semantic_memory": "All semantic memory results",
    "context": "Formatted memory context generated from search results",
    "totalResults": "Total number of episodic memory results",
    "apiResponse": "Raw response data from the MemMachine Search API"
  }

The AI Memory Node connects directly to the AI agent’s memory port in n8n workflows, enabling advanced memory operations and seamless integration with AI agent tools for managing conversational context.

Parameters

default template refers to the system’s built-in memory context template. Users may override it with a custom template.

Conversation History Output

• When enableMemoryTemplate is false (default, returns only chat history):

  {
    "chat_history": "Array of episodic memories (short-term and long-term), limited by contextWindowLength"
  }

• When enableMemoryTemplate is true (returns chat history with formatted context):

  {
    "chat_history": [
      {
        "content": "Formatted memory context"
      }
    ]
  }

MemMachine’s nodes include tracing capabilities for debugging and monitoring, which can be configured through node parameters.

Parameters

Jaeger UI

The following snapshot shows an example of tracing in the Jaeger UI.

• MemMachine n8n integration repository
• n8n Documentation
• n8n Community Nodes Installation
• MemMachine Documentation
• MemMachine API Documentation