For AI agents: a documentation index is available at the root level at /llms.txt and /llms-full.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
AcademyContact SalesHelp CenterDashboard
DocumentationAPI ReferenceIntegrationsAdministrationChangelog
DocumentationAPI ReferenceIntegrationsAdministrationChangelog
  • Get Started
    • Introduction
    • Aurora
  • Build
    • Create an Agent
    • The Agent Editor
    • Single-Prompt Agents
    • Memory
    • Knowledge Base
    • Variables
    • Version Control
  • Evaluate
    • Manual Testing
    • Custom Evaluations
    • Simulations
  • Launch
    • Deployment Options
      • Overview
      • Connect your Phone System
      • SIP Trunking Network ACL
      • Egress IP Addresses
      • Performance Metrics
      • Concurrency
      • Telephony Errors
      • Forward Calls to Synthflow
      • Call Transfer to SIP
        • How to Direct SIP Dialing
        • Viewing the SIP call ladder
        • Use Action Variables in Inbound SIP Calls
        • Forward Calls to the SIP Trunk
        • Forward Calls Using SIP Diversion Headers
        • Use the X-EI SIP Header
        • Use SIP X-Headers in Prompts, Actions & Transfers
      • Deploy in LATAM Regions
    • WhatsApp
    • Website and Apps
    • Launching a Chat Agent
    • Workflows
  • Learn
    • Analytics Dashboard
    • Logs
    • Export Call Data
  • Legal
    • Subscriber Terms
    • GTC - Direct Use (DACH)
    • GTC - Direct Use (US)
    • GTC - Distribution (DACH)
    • GTC - Distribution (US)
    • Business Associate Agreement
    • Privacy Policy
    • Imprint
    • AI Transparency Statement
    • Trust Vault
LogoLogo
AcademyContact SalesHelp CenterDashboard
On this page
  • General Flow:
  • Step 1: Create Custom Action
  • Step 2: Attach Action
LaunchTelephonyDial to the SIP Endpoint

Use action variables in inbound SIP calls

||View as Markdown|
Was this page helpful?
Previous

Forward calls to the SIP trunk

Next
Built with

General Flow:

Action Variables allow dynamic data to be fetched and injected into prompts during a call. This is done by enriching the call context before it reaches the agent, using an external API.

Step 1: Create Custom Action

How to create actions

Create an API-based action that fetches variables needed in your prompt.

  • The API must use POST.
  • Set run_action_before_call_startto true
  • It must accept an <external_id> in the request body.
  • It must return a JSON object where the keys match the variable names in your prompt.

Example: If your prompt uses a variable name, your API should return:

Passing Variables
{
"name": "John"
}

Note: No changes are needed for the prompts—they will work with the current variables. The only important point is to ensure that the same variable names used in the prompts are also used when returning values from the API.


Step 2: Attach Action

How to attach actions

Once the action is created:

  • Attach it to your agent.
  • The agent will automatically invoke this action before the call starts, enriching the prompt with the fetched variables.