AI Conversational Agent with Memory & Tools (OpenAI Assistant)

Unlock Advanced Conversational AI with this AI Agent

This n8n workflow empowers you to build a sophisticated AI Agent using OpenAI's powerful Assistants API. It's designed for creating chatbots or virtual assistants that can hold context-aware conversations, remember past interactions within a session, and even utilize external tools (like the included Calculator example) to provide more comprehensive responses. This AI Agent is perfect for automating customer interactions, providing intelligent support, or creating interactive product demos.

Key Features & Benefits

• OpenAI Assistant Integration: Directly leverages the capabilities of OpenAI Assistants, allowing for complex instruction following, knowledge retrieval, and the use of Assistant-native tools (like Code Interpreter or Knowledge Retrieval if configured in your Assistant) as well as Langchain tools.
• Persistent Conversational Memory: Utilizes Langchain's memory management nodes (Chat Memory Manager, Window Buffer Memory) to maintain conversation history, enabling natural, multi-turn dialogues where the AI remembers what was said earlier in the same session.
• Extensible Tool Usage: Comes with a Langchain Calculator tool example. You can easily add more Langchain tools to the OpenAI Assistant node, enabling your AI Agent to perform diverse actions, access external data, or interact with other services.
• Context-Aware Responses: By accessing and understanding the conversation history, the AI Agent provides more relevant, coherent, and helpful answers.
• Streamlined Setup & Testing: Features an n8n Chat Trigger for immediate interaction and testing of your conversational AI agent. Publicly accessible by default for easy demos.
• Versatile Applications: Ideal for building AI-driven solutions for B2C customer service, B2B client support, interactive SaaS onboarding, internal helpdesks, and more.

• An n8n instance (Cloud or self-hosted).
• An OpenAI API Key with access to the Assistants API.
• An existing OpenAI Assistant ID. You must create an Assistant in your OpenAI account (platform.openai.com/assistants). This Assistant can be configured with specific instructions, models, and OpenAI-native tools (like Code Interpreter or Knowledge Retrieval). The workflow's Calculator tool demonstrates a Langchain-based tool.

1. Download the n8n workflow JSON file.
2. Import the workflow into your n8n instance.
3. In the 'OpenAI Assistant' node: a. Select or create your OpenAI API credentials by clicking on the 'Credential To Connect' field. b. Critically, replace the placeholder 'Assistant ID' (e.g., asst_HDSAnzsp4WqY4UC1iI9auH5z) with your own OpenAI Assistant ID. You can find or create this in your OpenAI platform dashboard under 'Assistants'.
4. The 'Calculator' node is connected as an example Langchain tool. You can add more Langchain-compatible tools by connecting them to the 'Tool' input of the 'OpenAI Assistant' node. Ensure your OpenAI Assistant is configured to expect function calls if you want it to invoke these tools.
5. The 'Chat Trigger' node initiates the conversation. It's configured to load previous session messages using the connected 'Window Buffer Memory' node.
6. Review the 'Window Buffer Memory' node settings. The Session Key expression helps differentiate user sessions (defaults to using sessionId from the Chat Trigger). The Context Window Length determines how many past messages (user and AI turns) are kept in active memory for the conversation.
7. Activate the workflow (toggle the 'Active' switch to ON).
8. Use the 'Chat' button on the 'Chat Trigger' node (or its public URL if enabled) to interact with your AI Assistant. A sticky note in the workflow provides a simple test: tell the AI your name, then ask it what your name is in a subsequent message to test its memory.