Step 1: Install or Open n8n Sign up for n8n Cloud, or self-host by running the n8n Docker container locally. Once open, you land on the workflow canvas where you build by adding and connecting nodes. Self-hosting on n8n is a great option if you want your data to stay on your own infrastructure.

Step 2: Add an AI Agent Node Drag an “AI Agent” node onto the canvas. This is the brain of your workflow. Inside its settings, connect a language model by choosing a provider like OpenAI or Anthropic and entering your API key, so the agent can generate responses and make decisions.

Step 3: Give the Agent Tools Tools let your agent take real action. Add a “Tools Agent” setup and attach tools such as a web search node, a calculator, or a node that queries your own files. When you might ask the agent to “find today’s news about AI,” it autonomously calls the search tool to gather information.

Step 4: Add Memory for Context For more helpful agents, add a memory node using a vector store or window buffer, so the agent remembers earlier parts of the conversation. This matters if you build a customer-support agent that should track the context of a long back-and-forth.

Step 5: Connect Your Own Data with RAG To make the agent answer from your own documents, load files into a vector store connected to a Retriever node. Ask the agent only to use that retrieved content, and it will give grounded answers specific to your knowledge base instead of making things up.

Step 6: Add a Trigger and Test Place a trigger node, such as a Webhook or a scheduled interval, at the start of your workflow so it runs automatically. Press execute, send a test message, and watch each node run. Debug on the canvas, then activate the workflow. With n8n, you have just built a real AI agent that can automate meaningful work for you.