How to Integrate Rasa with WhatsApp

Set Up WhatsApp Business API

• Sign up for the Twilio WhatsApp Business API.

 

• After signing up, you must configure your phone number and create a Twilio project. Twilio provides a sandbox environment to test WhatsApp integration.

 

• Follow the instructions in the Twilio Console to link your WhatsApp business number and set up the necessary webhook URLs.

Install Rasa

• Ensure that Python and pip are installed on your system. Then, install Rasa using the following command:

 

pip install rasa

 

• Verify the installation by checking Rasa's version:

 

rasa --version

Create a Rasa Project

• Initialize a new Rasa project by running the following command and following the prompts:

 

rasa init

 

• This command creates a Rasa project structure in your directory with starter files configuration, training data, and more.

Configure Rasa for Twilio

• In your Rasa project directory, locate the `credentials.yml` file. This is where you will add your Twilio credentials.

 

• Add the following entries to `credentials.yml` to configure Twilio, replacing placeholder values with your Twilio account details:

 

twilio:
  account_sid: "your_twilio_account_sid"
  auth_token: "your_twilio_auth_token"
  twilio_number: "your_twilio_whatsapp_number"

Update Your Rasa Domain

• Open the `domain.yml` file in your Rasa project. Make sure you define the intents and responses necessary for your bot to handle.

 

• Update the `responses` section to include the messages you want to send back to users over WhatsApp.

Implement Twilio Webhook

• Create a new Python file in your Rasa directory called `twilio_connector.py`, which will manage messages between Rasa and Twilio.

 

• Within the `twilio_connector.py`, implement custom logic for Twilio webhook integration. An example snippet to get started:

 

from twilio.rest import Client
from flask import Flask, request
from twilio.twiml.messaging_response import MessagingResponse

app = Flask(__name__)

@app.route('/webhook', methods=['POST'])
def whatsapp():
    incoming_msg = request.values.get('Body', '').lower()
    # Define your bot's logic here
    resp = MessagingResponse()
    reply = resp.message()
    reply.body('Hello from Rasa with WhatsApp!')
    return str(resp)

 

• Make sure your Flask app is running when testing the bot with Twilio.

Run Rasa and Test the Integration

• Start your Rasa server using the following command:

 

rasa run --enable-api

 

• In a separate terminal, start the action server if you have custom actions:

 

rasa run actions

 

• Use a tool like ngrok to expose your localhost webhook to the internet for Twilio to access:

 

ngrok http 5005

 

• Update your Twilio Console webhook URL with the ngrok provided URL.

 

• Send a message to your Twilio WhatsApp number from WhatsApp, and observe the interaction with your Rasa bot.

Troubleshooting and Optimization

• If messages aren't flowing through, check your endpoint configurations in Twilio Console and ensure your credentials are correct.

 

• Use logs from Rasa and your Twilio webhook to debug any issues.

 

• Optimize response times by ensuring the server hosting your Rasa instance is sufficiently provisioned for your expected load.