Skip to main content

AI Chatbots

AI Chatbots lets you build an AI-powered chatbot for your website and customer conversations. You can configure how the chatbot behaves, customize its appearance, add a knowledge base, embed it on your website, and optionally connect messaging channels.
The workflow shown in this guide contains five setup steps: Configure, Customize, Train, Embed, and Channels.

How AI Chatbots works

The chatbot setup follows this workflow:
1

Configure

Define the chatbot’s name, instructions, messages, AI model, interaction type, and interface language.
2

Customize

Customize the chatbot’s logo, avatar, colors, branding, message timestamps, and widget position.
3

Train

Add knowledge sources such as URLs, PDFs, text, or Q&A so the chatbot can answer using your information.
4

Embed

Get the chatbot embed code, adjust its dimensions, and copy the code for use on your website.
5

Channels

Optionally connect supported messaging channels such as Telegram and WhatsApp.

Open AI Chatbots

From the Klyra sidebar, open AI Chatbots. The AI Chatbots dashboard gives you an overview of your chatbot activity and provides shortcuts for common tasks.
Klyra AI Chatbots dashboard showing chatbot metrics, creation shortcuts, and an existing chatbot

Dashboard overview

The dashboard can show:
  • Total Chatbots
  • Total Conversations
  • Total Messages
  • Active Domains
  • Total Tokens Usage
It also provides shortcuts for:
  • Create New Chatbot
  • View Recent Conversations
  • Chatbot Knowledge Bases
  • Chatbot Analytics
Under My Chatbots, you can see configured chatbots and their active status.
The metrics and chatbot information shown in the dashboard are workspace data. Your values will differ depending on your chatbot activity.

Create a chatbot

Select Create New Chatbot to begin. The setup wizard contains five steps:
  1. Configure
  2. Customize
  3. Train
  4. Embed
  5. Channels

Step 1: Configure the chatbot

The Configure Chatbot step defines the chatbot’s core behavior and conversation settings.
Klyra AI Configure Chatbot step showing chatbot identity, instructions, messages, model, interaction type, and interface language

Chatbot Title

Enter the title used to identify the chatbot configuration.

AI Chatbot Name

Set the name displayed by the AI chatbot. The interface also provides a Hide Chatbot Name option if you do not want the chatbot name displayed.

Chatbot Instructions

Use Chatbot Instructions to describe how the chatbot should behave and respond. The interface provides an example instruction:
You are a helpful AI assistant…
Use this area to provide the instructions that define your chatbot’s intended behavior and use case.

Restrict to data source and instructions

The Restrict to data source and instructions option controls whether the AI should answer using only the connected data sources and the instructions you provide. When enabled, the interface explains that the AI will only answer using connected data sources and the specified instructions.
Enable this setting when you want the chatbot’s responses to stay closely tied to the information you provide through its connected sources and instructions.

Custom Fallback Message

Enable Custom Fallback Message if you want to define what the chatbot should say when it does not know the answer. The interface provides a field for entering the custom response. For example, you can use a concise message that tells the visitor that the requested information is not available.

Greeting Message

Set the Greeting Message shown when the chatbot starts a conversation. The screenshot uses:
Hey, what can I do for you today?

Bubble Message

Set the Bubble Message displayed before the chatbot is expanded. The interface also provides Hide Bubble Message. When enabled, the bubble message will not be shown.

User Message Placeholder

Set the text displayed inside the message input before the visitor types. The screenshot uses:
Type your message…

AI Model

Choose the AI model that powers the external chatbot. The screenshot shows: OpenAI | GPT 4o mini The available models in your account may differ.

Interaction Type

Choose how users interact with the chatbot. The screenshot shows: Only AI

Chatbot Interface Language

Set the language used for the chatbot interface, including its menu, buttons, and interface text. The screenshot shows: Auto Click Next to continue.

Step 2: Customize the chatbot

The Customize Styles step controls the chatbot’s visual appearance.
Klyra AI Customize Styles step showing logo, avatar, colors, branding, timestamp, and widget position settings

Main Header Logo

Upload your logo using Drag and drop your logo or Browse. You can also enable Hide main banner logo if you do not want the main header logo displayed.

AI Avatar

Select an avatar for the chatbot. The interface displays several avatar choices and a button to add another option. You can also enable Hide AI avatar logo if you do not want the avatar displayed.

Customize colors

The Colors section lets you customize the visual colors used by the chatbot widget. The settings shown include:
  • Header Background
  • Header Text
  • AI Background
  • AI Text
  • User Background
  • User Text
The interface displays a color selector and a corresponding color value for each setting.
Use colors that maintain clear contrast between the chatbot header, AI messages, and user messages.

Branding

The Branding section includes an option to: Hide “Powered by Klyra AI” badge on the chat widget You can also set a Footer Link. The screenshot shows: https://klyra.ai

Message timestamps

You can enable: Hide Date and Time of Messages When enabled, date and time information is hidden from messages.

Widget position

Choose where the chatbot widget should appear:
  • Left Side
  • Right Side
The screenshot shows Right Side selected. Click Next to continue.

Step 3: Add a Knowledge Base

The Add a Knowledge Base step lets you provide information that the chatbot can use to personalize its responses.
Klyra AI Add a Knowledge Base step showing URL, PDF, Text, Q&A, and Website URL training options

Select a data source

The interface shows these data-source options:
  • URL
  • PDF
  • Text
  • Q&A
  • YouTube marked Coming Soon

Train via URLs

When using the URL option, the interface provides a Website URL field and a Train via URLs button. Enter the website information you want the chatbot to learn from, then select Train via URLs.
The screenshot shows YouTube as Coming Soon, so this tutorial does not treat YouTube as an available training source.

Other knowledge sources

Depending on the source you choose, the knowledge base can also use:
  • PDF documents
  • Text
  • Q&A
Use the source that best matches the information you want the chatbot to reference. Click Next after adding your knowledge source.

Step 4: Embed and integrate

The Embed and Integrate step provides the code required to place your chatbot on a website.
Klyra AI Embed and Integrate step showing chatbot embed code, width, height, and copy controls

Review the embed code

Klyra generates an embed snippet for the configured chatbot. The code shown in the interface includes chatbot-specific attributes as well as:
  • iframe width
  • iframe height
  • language
Do not manually replace chatbot-specific identifiers in the generated snippet unless you know exactly what you are changing.

Set the height

Use the Height control to set the chatbot height. The screenshot shows: 740px

Set the width

Use the Width control to set the chatbot width. The screenshot shows: 440px Adjust the dimensions to suit the layout of your website.

Copy the embed code

Click Copy to Clipboard to copy the generated embed code. You can then place the code in the appropriate location on your website.
The generated code is specific to the chatbot you configured. Use the code generated by Klyra rather than creating a new embed snippet manually.
Click Next to continue.

Step 5: Connect chatbot channels

The Chatbot Channel step is optional and lets you extend the chatbot experience through supported messaging channels.
Klyra AI Chatbot Channel step showing Telegram and WhatsApp channel options

Select a messenger

The interface currently shows:
  • Telegram
  • WhatsApp
Select the messenger you want to configure. The connected channels are displayed in a table with:
  • Channel
  • Channel ID
  • Actions
Channel configuration is optional. You can use the website chatbot without configuring an additional messaging channel.

Complete the chatbot setup

Once the five-step setup is complete, your chatbot can be used according to the configuration you selected. The resulting experience can include:
  • A customized website chatbot widget
  • Responses based on your configured instructions
  • Responses based on your connected knowledge sources
  • Your selected AI model
  • Your chosen interface language
  • Optional Telegram or WhatsApp channel connections

Recommended setup workflow

1

Start with clear instructions

Define the chatbot’s purpose, audience, and expected behavior in Chatbot Instructions.
2

Control the information it uses

If appropriate for your use case, enable Restrict to data source and instructions and provide the relevant knowledge sources.
3

Customize the experience

Add your logo, choose an avatar, set colors, configure branding, and select the widget position.
4

Train with useful information

Add the website URLs, PDFs, text, or Q&A content that the chatbot should use.
5

Test the website experience

Use the generated embed code and verify the chatbot’s appearance and behavior on your website.
6

Add messaging channels if needed

Connect Telegram or WhatsApp when you want the chatbot available through those channels.

Best practices

Write specific chatbot instructions

Instead of a generic instruction, describe the chatbot’s role, audience, and expected behavior. For example:
  • Explain what the chatbot helps users with.
  • Define the type of answers it should provide.
  • Specify how it should respond when information is unavailable.
  • Keep responses aligned with your business use case.

Use a focused knowledge base

Add information that is relevant to the chatbot’s purpose. A focused knowledge base makes it easier to maintain the information the chatbot should reference.

Configure a useful fallback

A clear fallback message helps users understand what to do when the chatbot cannot answer a question.

Match the visual design to your website

Use your logo, brand colors, avatar, footer link, and widget position to make the chatbot fit naturally into your website.

Test before publishing the embed

After copying the generated embed code, test the chatbot on the target website. Check its appearance, message input, responses, and available interactions.

Troubleshooting

The chatbot is not answering from my website content

Review the Train step and confirm that the intended source was added and trained. If you want the chatbot to stay restricted to your supplied information, review Restrict to data source and instructions in the Configure step.

The chatbot does not look right on my website

Review:
  • Header logo
  • AI avatar
  • Colors
  • Branding
  • Widget position
  • Embed width
  • Embed height
The Embed step lets you adjust the generated widget dimensions.

I do not want the chatbot name displayed

Enable Hide Chatbot Name in the Configure step.

I do not want the message bubble displayed

Enable Hide Bubble Message in the Configure step.

I do not want message dates and times displayed

Enable Hide Date and Time of Messages in Customize Styles.

I want to connect Telegram or WhatsApp

Open the Channels step and select the supported messenger you want to configure.

Quick reference

You’re ready to build your chatbot

You now know how to:
  • Create an AI Chatbot
  • Configure its identity and behavior
  • Set instructions and fallback responses
  • Choose an AI model
  • Customize the chatbot’s appearance
  • Add logos and avatars
  • Configure chatbot colors and branding
  • Add knowledge from URLs, PDFs, text, and Q&A
  • Generate and copy website embed code
  • Adjust chatbot dimensions
  • Connect supported messaging channels