zep-memory-v3

n8n community node for Zep Cloud v3 Memory - AI Agent persistent memory with Context Block

Package Information

Downloads: 28 weekly / 232 monthly

Latest Version: 0.3.3

Author: Fabio Henrique

Available Nodes

Zep Memory v3

Use Zep Cloud v3 for AI agent memory with Context Block

Documentation

n8n-nodes-zep-memory-v3

Supercharge your AI Agents with Zep Cloud v3 Long-Term Memory.

This community node integrates Zep Cloud v3 directly into your n8n workflows, providing your AI Agents with intelligent, long-term memory that goes beyond simple chat history.

🚀 Key Features

🧠 Intelligent Context Block: Feeds your AI Agent with summarized user details (USER_SUMMARY) and extracted facts (FACTS), not just raw text.
👀 Visual Execution Logs: See exactly what's effectively being sent and received from Zep right in the n8n Execution sidebar.
📉 Context Window Control: Save costs and tokens by limiting the number of recent messages sent to the LLM (e.g., last 10), while relying on Zep's summary for long-term context.
⚡ Zero-Config Threading: Automatically handles thread creation and persistence based on your threadId.
🛡️ Error Handling: Built-in retry logic and friendly error messages for common API issues.

📦 Installation

Open your n8n instance.
Go to Settings > Community Nodes.
Select Install.
Enter the package name:
```
n8n-nodes-zep-memory-v3
```
Install and restart n8n.

🛠️ Configuration

1. Prerequisites

A Zep Cloud account.
An API Key from your Zep Project Settings.

2. Setup in n8n

Create a Zep Cloud API credential in n8n and paste your API Key.
Add the Zep Memory v3 node to your workflow.
Connect it between your Chat Trigger (or user input) and your AI Agent node.

Parameter	Description	Recommended Value
Thread ID	Unique identifier for the conversation session.	`{{ $json.sessionId }}`
User ID	Unique identifier for the user (optional, defaults to Thread ID).	`{{ $json.userId }}`
Context Window	Number of recent messages to inject as chat history.	`10`
Use Context Block	Whether to inject Zep's summarized context (Facts/Summary).	`True`

📊 Visual Execution Feedback

Debug your memory flow like a pro. This node writes detailed logs to the n8n Execution Sidebar:

🔵 input: Shows the parameters requesting memory (e.g., loadMemoryVariables).
🟢 output: Shows the exact payload returned to the AI Agent (Context string + Messages array).

No more guessing what your AI "knows"!

🧩 Output Structure

The node initializes the AI Agent's memory with an object containing:

{
  "context": "<USER_SUMMARY>User loves sci-fi...</USER_SUMMARY><FACTS>...</FACTS>",
  "chat_history": [
    {
      "role": "user",
      "content": "Tell me about space."
    },
    {
      "role": "assistant",
      "content": "Space is vast..."
    }
  ]
}

⚠️ Migration Notice

For Open Source Users: This node (v3+) is exclusively compatible with Zep Cloud v3. Support for self-hosted Zep Open Source has been discontinued to leverage the advanced features of the new v3 Cloud SDK.

🤝 Support & Contribution

Found a bug? Have a feature request?

Open an Issue on GitHub.
Contributions are welcome! details in the repo.

zep-memory-v3Install