Get started with XUNA AI Conversational AI

XUNA AI Conversational AI lets you create voice agents that handle real conversations. This guide walks you through creating an agent, configuring its behavior, and embedding it on your website — using whichever approach fits your workflow.

Dashboard
CLI
API

Go to xuna.ai and sign in to your account.

Create an agent

Click Create agent, enter a name, and select the Blank template.

Configure behavior

In the Agent tab, set the first message and system prompt:First message:

Hi, this is Alexis from <company name> support. How can I help you today?

System prompt:

You are a friendly and efficient virtual assistant for [Your Company Name]. Your role is to assist customers by answering questions about the company's products, services, and documentation.

Tasks:
- Answer Questions: Provide clear and concise answers based on the available information.
- Clarify Unclear Requests: Politely ask for more details if the customer's question is not clear.

Guidelines:
- Maintain a friendly and professional tone throughout the conversation.
- Aim to provide concise answers. Limit responses to a couple of sentences.

Add a knowledge base

In the Knowledge base tab, upload documents, paste FAQs, or add URLs. The agent uses these to answer questions accurately.

Choose a voice

Open the Voice tab and select from 5,000+ voices. Preview voices before saving.

Test the agent

Click Test AI agent to start a live conversation with your agent directly in the dashboard.

Configure analysis

In the Analysis tab, add evaluation criteria (for example, solved_user_inquiry) and data collection fields (for example, user_question). These power conversation analytics after deployment.

Embed on your website

Copy your agent ID from the dashboard and add the widget snippet to your HTML:

<xuna-ai-convai agent-id="your-agent-id"></xuna-ai-convai>
<script src="https://unpkg.com/@xuna-ai/convai-widget-embed" async type="text/javascript"></script>

Install the CLI

npm install -g @xuna-ai/cli

Initialize a project

xuna-ai agents init

Authenticate

xuna-ai auth login

Enter your API key when prompted. You can find your key in the XUNA AI dashboard.

Create an agent

xuna-ai agents add "My Assistant" --template assistant

Deploy the agent

xuna-ai agents push --agent "My Assistant"

Generate the embed widget

xuna-ai agents widget "My Assistant"

The command outputs the HTML snippet you need to embed your agent on any website.

Install the SDK

pip install xuna-ai python-dotenv

Set your API key

Create a .env file in your project root:

.env

XUNA_AI_API_KEY=your_api_key_here

Define a system prompt

prompt = """You are a friendly and efficient virtual assistant for [Your Company Name].
Your role is to assist customers by answering questions about the company's products,
services, and documentation. Maintain a professional tone and keep responses concise."""

Create the agent

from dotenv import load_dotenv
from xuna_ai.client import XunaAI
import os

load_dotenv()

client = XunaAI(api_key=os.getenv("XUNA_AI_API_KEY"))

response = client.conversational_ai.agents.create(
    name="My voice agent",
    tags=["test"],
    conversation_config={
        "tts": {
            "voice_id": "aMSt68OGf4xUZAnLpTU8",
            "model_id": "eleven_flash_v2"
        },
        "agent": {
            "first_message": "Hi, this is Rachel from [Your Company Name] support. How can I help you today?",
            "prompt": {"prompt": prompt}
        }
    }
)

print("Agent created with ID:", response.agent_id)

Embed the agent

Use the agent ID returned in the previous step to embed the widget on your website:

<xuna-ai-convai agent-id="your-agent-id"></xuna-ai-convai>
<script src="https://unpkg.com/@xuna-ai/convai-widget-embed" async type="text/javascript"></script>

Next steps

Knowledge base

Upload documents and URLs so your agent answers questions with accurate, grounded information.

Tools

Give your agent the ability to call external APIs, run functions, and take action during conversations.

Authentication

Restrict access to your agent and pass signed user data into conversations.

Conversation analysis

Define success criteria and extract structured data from every conversation.

​Next steps

Knowledge base

Tools

Authentication

Conversation analysis

Next steps