🌊 Build a No-Code AI Agent with Flowise

Flowise runs locally on your machine. Make sure you have Node.js 18+ installed (download from nodejs.org). Then open a terminal and run:

This installs Flowise globally. Takes 1-2 minutes.

In the same terminal, run:

You'll see a message like "Flowise Server Now Listening On 3000". Open your browser and go to:

You should see the Flowise dashboard. Keep this terminal window open while you work.

In the Flowise dashboard:

1. Click "+ Add New" button (top right)
2. Select "Chatflow"
3. You'll see a blank canvas — this is where you'll build your agent

This canvas is your workspace. Every component (LLM, memory, tools) is a node you drag in and connect.

The LLM is the brain of your agent. To add one:

1. Click the "+" button on the canvas
2. Search for "ChatGoogleGenerativeAI" (or ChatOpenAI if you prefer)
3. Drag it onto the canvas
4. Click the node and paste your Gemini API key in the credentials field
5. Set model to gemini-2.0-flash for speed

You've added your AI brain. But it needs memory and a conversation interface.

Without memory, the agent forgets everything after each message. To fix this:

1. Click "+" and search for "Buffer Memory"
2. Drag it onto the canvas
3. Leave the default settings — they work fine

Buffer Memory stores the entire conversation in RAM. For longer chats use "Buffer Window Memory" which only keeps the last N messages.

Now we need a chain to tie the LLM and memory together:

1. Click "+" and search "Conversation Chain"
2. Drag it to the canvas
3. Connect the LLM node's output to the Chain's "Chat Model" input (drag a line)
4. Connect the Memory node to the Chain's "Memory" input

The chain orchestrates everything — when a user sends a message, it loads memory, sends to LLM, and saves the response back to memory.

Click "Save" at the top (give it a name like "My First Agent"). Then click the chat icon in the top right.

A chat window opens. Try:

It works! The memory node remembers across messages.

Let's make the agent more powerful by giving it a tool. We'll add Serper (a Google search API with a free tier):

1. Sign up at serper.dev (free 2,500 searches)
2. In Flowise, replace your Conversation Chain with an "AgentExecutor" node
3. Add a "Serper API" tool node and paste your Serper key
4. Connect the tool to the AgentExecutor

Now ask "What's the latest news on AI?" — the agent will search the web before answering.

The best part: every Flowise chatflow is automatically an API. Click the "</> API Endpoint" button at the top right.

You'll see a Python/cURL snippet like:

You can call your AI agent from any app — Python, JavaScript, mobile, anywhere.