Importable n8n AI Agent Workflow Builder with GPT-4o Memory

Build complete n8n automations from chat—now with GPT-4o memory

This importable n8n “meta-agent” workflow turns your plain-English automation idea into a fully importable n8n workflow JSON. Powered by OpenAI GPT-4o and conversation memory, it streams results back in chat—so you can go from requirement to working automation faster.

What this workflow does

• Receives a message via an n8n Chat trigger and streams responses back to the chat interface.
• Sends your request to an AI agent configured to design n8n workflows, instructing it to output an importable workflow JSON.
• Uses OpenAI GPT-4o as the chat language model to generate architecture, prompts, and the workflow JSON.
• Maintains short-term context with a buffer memory window, letting the agent reference recent conversation turns.
• Returns a full package to chat, including: project overview, requirements, architecture, node list, system prompt, setup notes, and the complete n8n workflow JSON.

Use cases

• Design an n8n automation/AI agent from a quick prompt like “Create a support triage agent that drafts replies and routes tickets.”
• Iterate on workflow requirements in a back-and-forth chat, with the agent preserving recent context using memory.
• Standardize how your team produces n8n workflows by generating consistent node lists, setup notes, and importable JSON.

Technical details

• Chat interface: n8nn8n-nodes-langchainchat trigger
• Agent design logic: n8nn8n-nodes-langchainagent
• LLM: n8nn8n-nodes-langchainlm chat open ai (configured for GPT-4o)
• Memory: n8nn8n-nodes-langchainmemory buffer window

Setup: Connect an OpenAI API credential (or update the existing one) for the GPT-4o node, then use the Chat trigger’s test/production chat URL in your n8n chat interface/embedding setup. Optionally adjust system instructions, max iterations, model name, and memory window size to match your defaults and output format.