Agent Configuration

Linking the components together and activating your Voice Agent

Agent Configuration

This section ties everything together: connect your agent to tenants, set limits, and deploy.

Step 1: Set Up Tenant Configuration

Tenants are your business entities. Use MongoDB to store:

  • Tenant ID (e.g., "your-business-id")
  • Widget keys for embedding on your site (must match origin URL for security).

Example tenant object snippet:

{
  "tenantId": "your-business-id",
  "widgetKeys": [
    {
      "key": "your_widget_key",
      "origin": "https://your-site.com"
    }
  ],
  "limits": {
    "maxConcurrentCalls": 15,
    "maxMonthlyMinutes": 2000
  }
}

Step 2: Connect Agents and Tools

Link your prompt and tools to the tenant. Specify agent repo if using GitHub for advanced setups (optional for starters).

Step 3: Manage Limits and Security

Set basics like max sessions or daily tokens to prevent abuse. (e.g., 10 concurrent sessions). Full rate limit details later—focus on defaults for now.

Step 4: Deploy and Test

  • Embed the widget on your site with the key.
  • Start the agent and test voice interactions.
  • Monitor sessions for issues.

Tip: For security, use secrets in tools for API auth. Try a demo at strategicmachines.ai to see configured agents in action.