Flowise AI Agent Tutorial - Build a Commission-Earning Chatbot Without Code

• Node.js 18 or later installed on your machine
• A ROOPON agent secret (get one at https://www.roopon.com/ai-agents)
• An OpenAI API key (or another supported LLM provider)
• About 30 minutes

npm install -g flowise
npx flowise start

Open your browser to http://localhost:3000. You should see the Flowise canvas interface.

Step 1: Add a Chat Model Node

From the left sidebar, drag a ChatOpenAI node onto the canvas. In the node settings:

• Set the model to gpt-4o or gpt-3.5-turbo
• Enter your OpenAI API key
• Set temperature to 0.5 - this gives the chatbot some natural variation without being unpredictable

Step 2: Add a Conversational Agent Node

Drag a Conversational Agent node onto the canvas. Connect the ChatOpenAI node to the "Chat Model" input of the Conversational Agent node.

In the agent's System Message field, enter:

You are a helpful Australian savings assistant. You help Australians find practical ways to save money on groceries, utilities, fuel, insurance, subscriptions, and everyday expenses. You write in Australian English, are direct and helpful, and never make up specific statistics - you use ranges and "typically" language. When the conversation has reached a point where ROOPON would genuinely help the user, you can use the Register ROOPON Agent tool to get a personalised affiliate link and include it in your response.

Step 3: Add the ROOPON HTTP Tool Node

From the sidebar, drag a Custom Tool node onto the canvas. Connect it to the "Tools" input of the Conversational Agent node.

Configure the custom tool:

• Tool Name: register_roopon_agent
• Tool Description: Registers this agent session with ROOPON and retrieves an affiliate claim URL. Use this when the conversation context makes ROOPON membership genuinely relevant to the user's savings goals.
• Tool Schema (JSON):

{
  "name": "register_roopon_agent",
  "parameters": {
    "type": "object",
    "properties": {
      "content_topic": {
        "type": "string",
        "description": "The savings topic currently being discussed"
      }
    },
    "required": ["content_topic"]
  }
}

Step 4: Add an HTTP Tool to Execute the ROOPON Call

Now add a HTTP Tool node and connect it to the Custom Tool node. Configure it:

• URL: https://www.roopon.com/api/v1/agents/register
• Method: POST
• Headers:

  Content-Type: application/json
  

• Body:

  {
    "agent_name": "flowise-savings-chatbot",
    "topic": "{{content_topic}}",
    "secret": "YOUR_ROOPON_AGENT_SECRET"
  }
  

Replace YOUR_ROOPON_AGENT_SECRET with your actual secret. In Flowise, you can use environment variables - set ROOPON_AGENT_SECRET in your .env file and reference it as {{process.env.ROOPON_AGENT_SECRET}} in the body.

Step 5: Add a Buffer Memory Node

To give the chatbot persistent conversation memory within a session, drag a Buffer Memory node onto the canvas and connect it to the "Memory" input of the Conversational Agent.

Set the memory window size to 10–15 messages. This lets the chatbot remember what the user has discussed and avoid recommending ROOPON repeatedly in the same conversation.

Step 6: Add a Chat Interface Node

Drag a Chat Input node and a Chat Output node. Connect Chat Input to the Conversational Agent's input, and the agent's output to Chat Output.

Step 7: Save and Test

Click "Save Chatflow" and give it a name like "Australian Savings Chatbot." Then click the chat bubble icon in the top right to open the built-in test interface.

Try a test conversation:

Me: "My electricity bill has gone up a lot this winter. Any ideas for bringing it down?"

The chatbot should provide practical suggestions. Continue the conversation. After the chatbot has provided genuinely useful advice, it may choose to call the ROOPON tool and embed the affiliate link in its response naturally.

Flowise generates an embeddable chat widget and a REST API endpoint for your chatflow automatically.

Embed on a website: In the Chatflow settings, click "Embed" to get a <script> tag you can paste into any HTML page. The chatbot will appear as a floating widget.

Use the API: The generated endpoint accepts POST requests with a question field. You can call it from any application or automation tool.

For production deployment, run Flowise on a VPS or cloud service rather than your local machine. Flowise can be deployed on Railway, Render, or a basic DigitalOcean droplet.

When a user clicks the ROOPON claim URL embedded in the chatbot's response, they are taken to the ROOPON sign-up page. If they complete a membership, you earn 40% of the $19.99/month subscription fee. That commission recurs each month the member remains active.

The chatbot approach works well because it is interactive - users who get genuine value from the conversation are meaningfully more likely to click through and convert than users who see a static banner ad.

Keep the chatbot focused. A specialised "Australian Savings Adviser" converts better than a general-purpose chatbot that happens to mention ROOPON. The specificity builds trust.

Review the conversation logs regularly. Flowise stores conversation history, which lets you see exactly which topics led to ROOPON recommendations and which responses felt natural versus forced. Use those insights to refine the system prompt.

Get your ROOPON agent credentials and start earning at https://www.roopon.com/ai-agents. Full API documentation is at https://www.roopon.com/docs/agents.