Turn live chat questions into clear AI replies that follow a strict JSON format. Great for support teams that need fast answers and clean data they can send to other systems. It speeds up replies and keeps the structure consistent across every message.

A chat event starts the flow and passes the text into an LLM chain. The chain uses the Ollama model to produce a JSON object with two fields called Prompt and Response. A mapping step then normalizes the output into a clean object, and a final set node controls what the user sees. If anything fails, an error branch returns a safe fallback so the chat never goes silent. You get reliable structure, simple routing, and outputs ready for storage or follow up tasks.

To set it up, you need n8n and an Ollama server with the llama3.2 model available. Add your Ollama connection in n8n, keep the provided prompt template, and confirm the mapping fields match your needs. Teams can expect faster answers and fewer mistakes, often cutting reply prep from minutes to seconds. Good fits include help desk chat, FAQ assistants, and any tool that needs structured AI responses for logging or downstream processing.

• Chat event trigger captures user messages the moment they arrive.
• Basic LLM Chain applies a clear prompt and requests strict JSON with Prompt and Response fields.
• Ollama model node runs llama3.2 to generate fast, local AI replies.
• JSON to Object mapping cleans and normalizes fields for downstream use.
• Structured Response node shapes the final payload returned to the chat.
• Error Response branch returns a default message if the chain fails.

• Reduce reply drafting time from 5 minutes to 30 seconds
• Automate 90 percent of response formatting with strict JSON
• Eliminate formatting errors by enforcing a fixed schema
• Keep support running with a safe fallback when the model fails
• Prepare data for CRMs or logs without extra cleanup

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Ollama. See the Tools Required section above for links to create accounts with these services.
3. Install and run Ollama on your server or local machine, then download the llama3.2 model so it is available for requests.
4. Open the Ollama Model node in n8n. In the Credential to connect with dropdown, click Create new credential and follow the on screen steps. Use your Ollama base URL, for example http://localhost:11434.
5. Open the Basic LLM Chain node and review the prompt. It should instruct the model to return a JSON object with Prompt and Response. Keep this strict so outputs stay consistent.
6. Check the connection from the Chat Trigger to the Basic LLM Chain. Confirm the field chatInput is used as the question input.
7. Open the JSON to Object node and verify manual mapping. Make sure the final object includes the exact keys you want returned to users or downstream apps.
8. Review the Structured Response node. Confirm it exposes only the fields you want visible in the chat reply.
9. Test the flow using the chat interface in n8n. Send a simple message and verify the reply contains Prompt and Response in valid JSON.
10. If you see errors, check the Error Response node is connected to the chain error output. Ensure it sends a helpful fallback message.
11. Troubleshoot common issues: if the model is not found, download the model in Ollama. If replies are empty, confirm the base URL, credentials, and that the prompt requests strict JSON.