> ## Documentation Index
> Fetch the complete documentation index at: https://docs.xuna.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Deploy your conversational AI agent

> Choose how to deploy your XUNA AI voice agent — from a no-code website widget to native mobile SDKs, phone integration, and a raw WebSocket API.

Once you've configured your agent, you need to get it in front of users. XUNA AI Conversational AI supports several deployment paths, from a single HTML snippet to full telephony infrastructure. Pick the approach that fits your stack and use case.

<CardGroup cols={2}>
  <Card title="Website widget" icon="browser" href="/deploy/widget">
    Drop two lines of HTML into any webpage to embed a fully functional voice agent with no additional code.
  </Card>

  <Card title="React SDK" icon="react" href="/deploy/react-sdk">
    Use the `@xuna-ai/react` package to build custom voice interfaces with hooks, controls, and client tools.
  </Card>

  <Card title="Mobile SDKs" icon="mobile" href="/deploy/mobile-sdks">
    Integrate voice agents into iOS (Swift), Android (Kotlin), or cross-platform React Native apps.
  </Card>

  <Card title="Phone & telephony" icon="phone" href="/deploy/phone-telephony">
    Connect a Twilio number to handle inbound calls, trigger outbound campaigns, or route via SIP trunk.
  </Card>

  <Card title="WebSocket API" icon="plug" href="/deploy/websocket-api">
    Build a fully custom integration using the low-level WebSocket protocol for audio streaming and conversation control.
  </Card>
</CardGroup>

## Choosing a deployment method

| Method            | Best for                                    | Auth required      |
| ----------------- | ------------------------------------------- | ------------------ |
| Widget            | Quick website embeds                        | No (public agents) |
| React SDK         | Custom web UIs                              | Optional           |
| Mobile SDKs       | iOS / Android apps                          | Optional           |
| Phone & telephony | Call centers, outbound campaigns            | Yes                |
| WebSocket API     | Custom servers, hardware, non-web platforms | Optional           |

<Info>
  All deployment methods use the same underlying agent configuration. Change your agent's prompt, voice, or tools in the dashboard and every deployment picks up the update automatically.
</Info>

## Authentication and private agents

Public agents can be accessed directly with an `agentId`. If your agent uses sensitive prompts or private data, make it private and generate a signed URL or conversation token on your server before passing it to the client. See [Authentication](/configure/authentication) for details.