Hello World with Pro Config

Key Concepts in This Chapter: - Render Types - Inputs/Outputs

"Hello World" Example

We start by building a simple agent that can render and output “Hello World” with Pro Config. Here is what the config looks like:

For better readability of this code block, you could copy and paste this into any code editor app like VS Code and select JSON as its format.

Whenever you upload a new Pro Config, you can click the ≡ button on the bottom left and click "Clear Memory" button to reset your agent.

{
  "type": "automata",
  "id": "hello_demo",
  "initial": "home_page_state",
  "inputs": {},
  "outputs": {},
  "transitions": {},
  "states": {
    "home_page_state": {
      "inputs": {},
      "tasks": [],
      "outputs": {},
      "render": {
        "text": "Hello World! Welcome to this demo. Click 'Start' to chat!",
        "buttons": [
          {
            "content": "Start",
            "description": "Click to Start.",
            "on_click": "start_demo"
          }
        ]
      },
      "transitions": {
        "start_demo": "home_page_state"
      }
    }
  }
}

Above is the syntax with all the available fields for a Pro Config JSON file. Using the code above should allow you to do the following interaction with your agent.

Hello world using pro config

In Pro Config, each Automata can be defined using a JSON file. An Automatahas a unified id and several states as its children. In the simple example above, the Automatahas only one state called home_page_state which simply outputs an intro message. It also has a button called Start which can jump to home_page_state itself like an iteration (we will get to more complicated transitions in the future.)

AtomicState

An atomic state is a state executing real tasks, which usually are small functional modules, such as LLM module and TTS module.

These are usually embedded as a part of a bigger automata.

Automata

Automata shares many fields with AtomicState. They differ by the following:

• AtomicState has the lack of properties.is_chat_allowed and tasks fields.

• The different special events it can handle.

• initial , states and context fields.

Put in simpler words, one Automata can contain different AtomicState. Your Pro Config should be of the type Automata and tasks should be AtomicState.

For more information about the difference between AtomicState and Automata, please refer to the dev doc.

Defining Inputs/Outputs

The states of an Automata is a dictionary of AtomicState . An AtomicState take inputs from the user, process some tasks, return outputs, and render some content (such as text/image/button). In the example above, we rendered a text message and a button. We will now demonstrate how to add inputs and outputs to an Automata

{
  "type": "automata",
  "id": "hello_demo",
  "initial": "home_page_state",
  "inputs": {},
  "outputs": {},
  "transitions": {},
  "states": {
    "home_page_state": {
      "inputs": {
        "intro_message": {
          "type": "text",
          "user_input": true,
          "default_value": "Hi, this is your Pro Config Tutorial Agent"
        },
        "tts_widget_url": {
          "type": "text",
          "user_input": true,
          "default_value": "https://app.myshell.ai/widget/mEjUNr"
        }
      },
      "tasks": [],
      "outputs": {
        "intro_message": "{{intro_message}}",
        "voice_id": "{{tts_widget_url}}"
      },
      "render": {
        "text": "Hello Word! Welcome to this demo. Click 'Start' to chat!",
        "buttons": [
          {
            "content": "Start",
            "description": "Click to Start.",
            "on_click": "start_demo"
          }
        ]
      },
      "transitions": {
        "start_demo": "home_page_state"
      }
    }
  }
}

In the example above, we are doing the following:

• Lines 7-18: Gathering inputs intro_message and tts_widget_url from the user. The type is text which means that the user will be prompted to input data (unlike IM where user needs to type in the chat to provide an input). If the user_input property is false, the user will not be prompted to input via a form, and a new variable with the value of default_value will be automatically generated.

Input fields prompted to user when input type text is used

• Lines 19-22: The output section in pro config runs after all the tasks are completed for the automata or atomic state. In this section, you can create or manipulate variables. This is also the place you can manipulate any variables you create in context. If you want to store output for the task performed, this is the way. We use an expression wrapped by double curly braces {{expression}} to assign the value of an output variable. The expression should be written in JavaScript, adhering to the ECMAScript 5.1 standard, as we currently support only this version. In the example, we are using this section to save the inputs to variables.

Now, you have learned how to build a basic app with inputs and outputs. In the next chapter, we will learn how to build a workflow to achieve more complicated functionalities empowered by AI Modules.