Embedding

In addition to using DataBot in a dedicated webpage—either with just the chatbot or with both the chatbot and dashboards—you can also embed it directly into your own website or application, as follows:

Embedding in a Webpage or App

You can embed DataBot anywhere using an iframe. Make sure to include allow="microphone" so the voice button works.

<iframe src="https://..." allow="microphone"></iframe>

Single Sign-On (SSO)

With SSO, you can authenticate users within your own app or website—eliminating the need to use DataBot’s built-in login. Once users log in to your platform, you can seamlessly embed DataBot in an iframe, allowing them to access it directly within your app or website.

You can control user access levels by specifying the group_memberships and custom_user_attribute_assignments parameters.

• group_memberships defines which tables or views a user can access.
• custom_user_attribute_assignments is used to manage access at row level and also for dashboard cards—both when viewing dashboards and when using dashboard cards as data source.

From your backend, send a request to /sso-auth to receive an encoded user session ID (EUSI):

import jwt
import time
import requests

base_url = 'http...'  # the databot url. E.g., "http://localhost:5000"  
DATABOT_AUTH_KEY = ***  # this is the environment variable you provided 
                        # when deploying DataBot

# Generate JWT token
jwt_payload = {
    "email": "user@example.com",  # replace with the actual user email
    "iat": int(time.time()),
    "exp": int(time.time()) + 10,  # expires in 10 seconds

    # optional
    "group_memberships": ['group1', 'group2'],  # example values
    "custom_user_attribute_assignments": {  
        "attr1": "12345",  # example key-value 
        "attr2": "abcd",  # example key-value
    }
}

try:
    # Sign the JWT using HS256
    jwt_token = jwt.encode(jwt_payload, DATABOT_AUTH_KEY, algorithm="HS256")

    # Place token in Authorization header
    headers = {
        "Authorization": f"Bearer {jwt_token}",
        "Content-Type": "application/json"
    }

    response = requests.post(f"{base_url}/sso-auth", headers=headers)
    print(response.json())  
    # {'eusi': 'eyJhbGciOiJIUzI1NiI...'}
except Exception:
    print(response.text)
    # in production, add alerts to be notified about the error

You would then use the EUSI to build the iframe URL:

<!-- Replace [EUSI] with the actual session ID -->
<iframe src="https://your-databot-url/?eusi=[EUSI] allow="microphone"></iframe>

<!-- Example -->
<iframe src="https://your-databot-url/?eusi=eyJhbGci..." allow="microphone"></iframe>

You can use the environment variable SSO_SESSION_TIMEOUT_SECONDS to specify how many seconds the iframe URL should remain valid. The default is 15 seconds, based on the assumption that your backend generates a new EUSI on each page load.

Embedding in Salesforce

The easiest way to embed DataBot as an iframe inside Salesforce is using a Visualforce Page:

<apex:page showHeader="false" applyBodyTag="false" applyHtmlTag="false">
    <iframe 
        src="[iframe url]" 
        width="99%" 
        height="800" 
        allow="microphone"
        style="border: 1.5px groove #ccc;">
    </iframe>
</apex:page>