Private, Self-Hosted, Dynamic Ollama Chat Agent

Who Is This For? This workflow suits AI fans, developers, and users who care about privacy. It helps people who run Ollama on their own computers and want smart choices between different AI models. What Does It Do? This tool gets chat messages. It checks the message to pick the best local AI model from Ollama. It then uses that model to answer the user. Everything stays private on your machine. It keeps track of past chats for better talks. How to Set Up 1. Install and run Ollama on your computer. 2. Download needed models with Ollama commands, like ollama pull phi4. 3. Set up Ollama API in n8n at http://127.0.0.1:11434. 4. Turn on the workflow and chat with it. Requirements You need Ollama installed and running. Pull models like qwq, llama3.2, phi4, qwen2.5-coder:14b, granite3.2-vision, llama3.2-vision. Use n8n with LangChain nodes. Set credentials for local Ollama. How to Customize Add or remove models in the router's rules. Change system messages to fit your needs. Update the decision tree for your tasks. Add steps for special inputs like files. Key Technologies: n8n Workflow Automation,Ollama Local LLMs,LangChain Nodes (Agents, Memory, Chat Trigger),JSON Parsing for Dynamic Routing Automation Value: Keeps all data private by running AI locally, no cloud needed. Saves time by auto-picking the best AI model for each chat. Boosts AI answers with smart model choices for tasks like coding or image analysis. Cuts costs and speeds up responses with lightweight local models. Builds ongoing chats with memory for natural talks. Best Practices: Run all AI local with Ollama to guard privacy and avoid cloud fees. Use clear rules and examples in prompts for reliable model routing. Keep memory windows small to save resources in long chats. Test with varied prompts to tune the decision tree. Update models often via Ollama to get better AI performance.