n8n Website Support Chatbot with OpenAI GPT-5 Memory

Turn your website into a 24/7 OpenAI-powered support channel

This n8n workflow adds a website chat widget that routes visitor questions to an OpenAI-powered support agent with short-term conversation memory—so you can deliver consistent, friendly answers directly on your site.

What this workflow does

• Receives visitor messages from an embedded website chat widget via a public webhook.
• Generates support replies using an OpenAI chat model, following your configured support guidelines for tone and accuracy.
• Maintains recent context with a buffer window memory so responses remain consistent across multiple turns in the same conversation.
• Returns the final answer back to the chat widget as the visible assistant response.

Use cases

• SaaS support: Answer common onboarding, billing, and feature questions without waiting for a human agent.
• Pre-sales Q&A: Help visitors understand pricing, availability, and key product capabilities while they browse.
• Help center deflection: Guide users to the right next step using your support instructions and conversation context.
• Small teams: Provide reliable 24/7 website assistance without building a custom chatbot from scratch.

Technical details

• Chat Trigger: n8nn8n-nodes-langchainchat trigger to accept messages from the website widget via webhook.
• OpenAI Chat Model: n8nn8n-nodes-langchainlm chat open ai to produce friendly, concise support responses.
• Agent & Orchestration: n8nn8n-nodes-langchainagent to manage the support flow.
• Memory: n8nn8n-nodes-langchainmemory buffer window to keep short-term conversation context for consistent multi-turn answers.

Setup

• Add or update your OpenAI API credential for the chat model.
• Copy the chat widget/webhook URL from the trigger and embed/configure it on your website.
• Review and adjust assistant system instructions, widget title/subtitle, and allowed origins for your security requirements.