Skip to content
Nappai Documentation

Quick Start Guide

Welcome to NappAI! In this guide, we’ll walk you through building your first automation: a simple chatbot that answers your questions using Google’s AI.

What you will build

You will create a flow with four essential components (nodes) that receive your question, process it with an AI, and return a response in a chat interface.

Prerequisites

To follow this tutorial, you will only need:

  • An NappAI account.
  • A Google Generative AI API Key. You can get one for free here.

Building Your First Project

Step 1: Create an Empty Project

Let’s start from a blank canvas to understand each piece.

  1. On the main dashboard, click on + New Project. Botón de Proyectos

  2. Select the Empty Project template and then click on Use this template. Plantilla de proyecto vacío

  3. Perfect! You are now in the project view. Click on Add your first component to get started. Vista de proyectos

Step 2: Add Components to the Flow

Now let’s add the “bricks” of our automation. In NappAI, components are organized into categories so you can easily find what you need: Core, AI Nodes, Databases, etc.

Menú de componentes mostrando las categorías.

For this smart flow, we need the following components:

  1. Chat Input: Receives user messages.
  2. Agent: Acts as the “director” of the operation, coordinating the logic and the use of tools.
  3. Gemini: Google’s AI model, which will function as the “brain” the agent will use.
  4. Chat Output: Displays the final response in the interface.

To add them, simply find them in the right-hand menu and drag them one by one to the workspace.

Step 3: Connect the Components

The connection with an Agent is special. There is a main data flow and a connection for the model.

  1. Connect the main flow: Click on the output point of Chat Input and drag it to the Input entry of the Agent. Then, connect the Text output of the Agent to the input of the Chat Output.
  2. Connect the Model to the Agent: Click on the output point of the Gemini component and drag it to the input point called Model at the bottom of the Agent.

Your final flow should look exactly like this: Flujo completo con el Agente y Gemini conectados.

Step 4: Configure the Components

Now, let’s give instructions to each component. Click on each one to open its configuration.

  1. Chat Input: Needs no configuration. Easy!

  2. Gemini: This is where you configure access to the AI model.

    • Click on Create New credential.
    • Add a name for the credential.
    • Paste your Google API Key that you obtained in the prerequisites.
    • Once the credential is added, select the model (e.g., gemini-2.0-flash).

  3. Agent: This is where the magic happens! This is the main configuration.

    • Agent Name: You can give it a descriptive name, like ReactAgent.
    • System Prompt: This is the agent’s brain. It defines its personality and its main task. Paste the following text, which is very similar to the previous prompt: 
      You are a helpful and creative assistant. Answer the user's question clearly and concisely.
      

      User question(io: {{aporte}})

    Important: Note that the variable for the user’s input here is {{aporte}}. The Agent receives the information in this variable.

    Panel de configuración del componente Agente.

  4. Chat Output: Also needs no configuration.

Step 5: Test Your Chatbot!

You’ve reached the end! It’s time to see your creation in action.

On the right side of the screen, you will see the Test Environment which has a bug icon. Entorno_de_prueba. Press it and it will open the test space. Espacio de prueba. Type a question in the text field and press Enter. In a few seconds, the AI will respond to you.

Congratulations! You have built and tested your first AI workflow in NappAI.

What now? Next Steps

You have mastered the basics. Why not try something more?