n8n-nodes-supermemory

Give your n8n AI agents persistent, semantic memory with Supermemory. This package includes two nodes:

• Supermemory Memory — plugs into an AI Agent's memory slot so it remembers users across conversations
• Supermemory — 22 workflow operations for memories, documents, search, containers, conversations, and profiles (also usable as an AI Agent tool)

n8n is a fair-code licensed workflow automation platform.

Installation

Follow the n8n community node installation guide.

1. Go to Settings > Community Nodes
2. Select Install
3. Enter n8n-nodes-supermemory
4. Agree to the risks and select Install

Credentials

1. Get a free API key at console.supermemory.ai
2. In n8n, go to Credentials > New Credential
3. Search for Supermemory API
4. Enter your API key
5. (Optional) Change the Base URL for self-hosted instances

AI Agent Memory

The Supermemory Memory node gives any n8n AI Agent long-term, semantic memory. Unlike buffer memory that stores raw chat history, Supermemory automatically extracts and organizes what matters — the agent gets a compiled user profile with permanent facts and recent context.

How it works:

• On each turn: the conversation is sent to Supermemory, which extracts memories automatically
• Before each response: the agent receives the user's profile (static facts like name/preferences + dynamic context like recent topics)
• Across sessions: memories persist — the agent remembers the user next time

Setup:

1. Add an AI Agent node
2. Connect a Supermemory Memory node to its Memory input
3. Set Container Tag to a user identifier (e.g., {{ $json.userId }})
4. Set Session ID to a conversation identifier (defaults to execution ID)

That's it. The agent now has persistent memory.

AI Agent Tool

The main Supermemory node is also marked as usableAsTool, meaning you can connect it to an AI Agent's Tool input. This lets the agent explicitly decide when to store memories, search knowledge, or look up a user profile — useful when you want the agent to have more control over when and how it uses memory.

Workflow Node Operations

The Supermemory node can also be used as a regular workflow node with 22 operations:

Memory

| Operation | Description | |-----------|-------------| | Create | Create memory entries with optional static flag, expiry, and temporal context | | Delete | Forget a memory entry by ID or content match | | Get Many | List memory entries with pagination, sorting, and filters | | Search | Search memories with threshold, search mode (memories/hybrid/documents), reranking | | Update | Update a memory (creates a new version with full history) |

Example — Store a user preference:

1. Set Resource to Memory, Operation to Create
2. Set Container Tag to user-123
3. Set Content to "User prefers dark mode and metric units"

Document

| Operation | Description | |-----------|-------------| | Add | Add content by URL or text | | Batch Add | Add up to 600 documents at once | | Bulk Delete | Delete multiple documents by IDs | | Delete | Delete a single document by ID | | Get | Retrieve a document by ID | | Get Chunks | Get processed chunks of a document | | Get Many | List documents with pagination, sorting, and filters | | Get Processing | Get documents currently being processed | | Update | Update an existing document | | Upload File | Upload and process a binary file |

Example — Index a webpage:

1. Set Resource to Document, Operation to Add
2. Set Content to https://example.com/article

Container

| Operation | Description | |-----------|-------------| | Delete | Delete a container tag and all its contents | | Get Settings | Get settings for a container tag | | Merge | Merge two container tags into a target | | Update Settings | Update the entity context prompt |

Search

| Operation | Description | |-----------|-------------| | Search Documents | Semantic search (RAG) with chunk threshold, reranking, query rewriting | | Search Memories | Search memory entries with mode, threshold, and include options |

Conversation

| Operation | Description | |-----------|-------------| | Ingest | Send a conversation for automatic memory extraction |

Profile

| Operation | Description | |-----------|-------------| | Get | Get a compiled user profile with static and dynamic memories |

Compatibility

• Requires n8n version 1.0 or later
• Tested with Supermemory API v3/v4

License

MIT