Quick Start with PingPong

Follow this PingPong chatbot to get started with Framely.

This guide will walk you through using Framely platform to build and test a simple PingPong chatbot. When interacting with this chatbot, a user can get greeting message, and response of "pong" to his input message "ping":


Hi! I'm a virtual PingPong chatbot. How can I assist you today?




Pong. What else can I do for you?


No, thanks


Thank you for contacting the virtual PingPong chatbot. Have a great day!

After this simple PingPong chatbot, you can easily build level 1 conversational user interface, and easily manage how the messages are presented to users per your own business logic. For example, if you are a restaurant, you can let your users check your business hours as follows:


Hi! I'm a virtual Restaurant chatbot. How can I assist you today?


What is your hours?


We open every day from 10:00am to 9:00pm. Can I help you with anything else?


No, thanks


Thank you for contacting the virtual Restaurant chatbot. I am glad I could help you today. Good-bye!

Now, let's begin:

Before You Begin

Before you start follow this guide, it is suggested that you go through the following list first.

  1. Get familiar with the basic idea of building conversational apps.
  2. Make sure that you meet the prerequisite
  3. Sign Up For Framely Account

Create Chatbot

After logging in, there will be an org (short for organization) created automatically for you, which is how you manage everything on the Framely platform. To create chatbot, follow these steps:

  1. Select an org from your org label:

org list

  1. Create a chatbot :
    1. Click Chatbots in the left sidebar menu.
    2. Click the Create button on the right side.

    create chatbot

    1. Complete the form for basic chatbot settings, here you only need to care about the Project Label, Region and Add language. All can be modified after creation except the Region field:

      1. Enter your chatbot's name in the Project Label field, start with lowercase, for example pingpong.
      2. Select your preferred Region.
      3. Select the languages for your chatbot in the Add Language field.

      complete form

      Need To Know

      • When you create a chatbot, you must specify a region. For the best performance, you should choose a region that is near your services and end-users.

      • Once a chatbot is created, its Region cannot change. In order to change a chatbot's region, you must export or clone a new chatbot with a different region.

    2. Once you are done with the form, click the Save button.

      create save

Once created, you can see the chatbot shown like the following:

enter chatbot


Chatbot is essentially just a set of intents and its dependency. An intent offers conversational interface to some functionality, typically as part of some services. In conversation layer, the intent represents what a user wants, is typically expressed by full sentences or verb phrases in user utterances.

When you create a chatbot, the following Default Intents are created automatically for you, and you can modify them as desired.

Default Intents

  • Greeting: a default welcome intent, this intent has simple expression exemplars like "Hi" or "Hello" that are matched when the user begins a conversation with your bot. This intent could return a response to let the user know what your bot does or what they can say to begin a conversation.

  • Goodbye: a default ending intent, this intent returns a response to let the user know the conversation or service is ending soon.

  • Main: a default mechanism for each chatbot.

default intents

Create a New Intent

For each chatbot, you can define many intents. The steps in this section create an intent that can response with "pong" to message "ping".

1. Create an intent

  1. Click the Create button on the right side.

create intent

  1. Enter PingPong in the Intent Label field, press enter.

intent label


  • Label is not a name, it is identifier, a language independent aspect for semantics in Framely.
  • As a type label, Intent Label should start with capital case, limited 2-100 characters, and only support letters, digits and underscores.

2. Build PingPong intent

Separation of concerns is essential in increasing productivity and reducing the cost of building things. Framely decompose chatbot into 3 layers: schema, interaction and language perception. In this way, the exact conversational experience should be easily controlled by you based on arbitrary business logic. For example, it should be more easier to keep a multilingual chatbot with the same structure which can provide a consistent experience in each language. For more information, see Separation of Concerns .

In this PingPong bot, we can just care about two layers: interaction and language. The difference between interaction and language is whether the entry is language-dependent or not. Language-related aspect like expression, prompt, reply for dialog understanding and template for text generation needs to be defined on language level, while others should be on structure level.

So building PingPong intent, should add response on the structure level first:

  1. On STRUCT level, click the Response tab, add Simple Reply in the Default Action section.

add response

  1. Click Commit icon in the right sidebar, for propagating structure level instances to each language level.

commit pingpong struct


In this way, you can always keep all language bots with the same structure which can provide a consistent experience in multilingual chatbot. For more information, see Multi Language.

  1. Select language in the second topbar, switch to the language level. In this case, we will switch to EN (short for English) level.

switch pingpong en

Now let's add language-related aspects:

  1. On the language level, click the Expression tab, enter Ping Pong in the Alias section, press enter. This field is the display name of an intent, and it is also an example of what a user might type or express this intent.

pingpong alias

  1. In the Expression tab, enter ping in the Expressions section, press enter.

pingpong expression

  1. Click the Responses section, enter Pong in the Simple Reply section.

pingpong simple reply


Let's try this PingPong bot, which can handle a basic conversation with a welcome message and response with "pong" to message "ping". Framely provides a built-in test feature "Try It Now" to help you to uncover bugs with the debug info, you can test your chatbot by typing messages.


As Try It Now can only test committed content, please make sure both structure level and language level have been committed.

  1. Make sure you have committed on each level.

commit structFigure 1: click commit on STRUCT level

commit langFigure 2: click commit on EN level

  1. Click Try It Now icon in the right sidebar, the test field will slide out.

try it nowFigure 1: click Try It Now icon

try it nowFigure 2: Try It Now slide out

  1. Click Connect, and you can test your bot by typing messages.


  1. Enter ping in the text entry and press enter, the bot responds Pong.

pingpong test


You can restart your testing by clicking "RESET CONTEXT" icon, and create test cases by clicking "Save Test case" icon, then you can run test cases when needed.

Last Updated: