Skip to main content

Deploy as a Voice Agent

Give your agent a voice. SmythOS allows you to deploy your agent as a voice-enabled assistant in two primary ways: as an embeddable chat widget for your website, or as the backend for a custom Amazon Alexa Skill.

TL;DR

You can deploy your agent in two ways:

  1. Website Embed: Get the code snippet from Deploy → Voice.
  2. Alexa Skill: Get the Endpoint URL from the Deploy → Voice.

Embed a Voice Chat on Your Website

This option adds a voice-enabled chat widget to your website, allowing users to talk to your agent directly from their browser.

1. Find Your Code Snippet

  1. In your agent, navigate to the Deploy tab.
  2. In the Embed section, find the Voice option and click Get Endpoints.
    Finding the Voice deployment option in the SmythOS Deploy tab.

    Click 'Get Endpoints' on the Voice option in the Embed section.

  3. In the "Voice and Alexa Integration" modal, find the Voice Integration section and copy the Code Snippet.
    The SmythOS Voice and Alexa Integration modal.

    Copy the Code Snippet from the Voice Integration section.

2. Add the Code to Your Website

Paste the snippet you copied just before the closing </body> tag in your website's HTML file.

The snippet will look similar to this:

<script 
src="https://<your-id>.smyth.ai/voice.agent.js"
id="smythos-voice-embed"
data-voice-embed-id="https://<your-agent-id>/voiceEmbedment.init?cid=..."
>
</script>
<script>
voiceEmbedment.init()
</script>

3. How it Looks

Once embedded, a microphone icon will appear on your site. When activated, the voice interface provides clear visual feedback and controls for the user.

User interface of the SmythOS voice chat embed.

The voice chat interface showing active listening/speaking status with microphone controls.

Personal Information

Note that to use the voice embedment feature, you may need to provide personal details for compliance and security purposes.

Connect to an Amazon Alexa Skill

This option uses SmythOS to provide the backend HTTPS endpoint for a custom Amazon Alexa Skill.

Step 1 (Development): Get Your Test Endpoint

The fastest way to start developing your Alexa skill is to use the Test panel. This gives you a development endpoint without needing to run a full deployment.

  1. In your agent, click the Test button in the top-right debug panel.
  2. Select the Voice tab.
  3. Copy the Alexa Dev Endpoint. This is the URL you will use for initial setup and testing in the Alexa Developer Console.
Finding the Alexa endpoints in the Test with Voice sidebar.

Use the Test panel's Voice tab to get your development endpoint.

Step 2 (Production): Get Your Stable Production Endpoint

Once you have tested your skill and are ready to deploy it, you should switch to the permanent, stable endpoint.

  1. Gi to the Deploy tab.
  2. Click Get Endpoints on the Voice option.
  3. From the "Voice and Alexa Integration" modal, copy the Alexa Prod URL (e.g., ...agent.pstage.smyth.ai/alexa).
The SmythOS Voice and Alexa Integration modal showing production URLs.

Use the Alexa Prod URL from the Deploy modal for your live skill.

Step 3: Configure Your Skill

Now that you have your endpoint URL (from Step 1 or 2), you are ready to configure your skill.

Next Steps for Alexa

Once you have this URL, you are ready to create your skill. Go to our detailed guide on Alexa Skill which will walk you through the entire process in the Alexa Developer Console.

What’s Next?

Now that you have your voice endpoints, you can: