Synthflow icon

Synthflow

Interact with Synthflow AI API
Join our community

Actions8

Overview

This node interacts with the Synthflow AI API to manage voice agents and calls. Specifically, the 'Create Agent' operation allows users to create a new Synthflow voice agent of types outbound, inbound, or widget. The node is useful for automating the creation and configuration of AI-powered voice agents that can handle calls, including setting prompts, greetings, language models, voice IDs, and additional agent-specific settings. Practical applications include setting up customer service agents, sales call agents, or interactive voice response systems.

Use Case Examples

  1. Creating an outbound customer service agent with a specific greeting and prompt.
  2. Setting up an inbound voice agent with consent and recording enabled for compliance.
  3. Configuring a widget agent for website interaction with a custom language model and voice.

Properties

Name Meaning
Agent Type The type of the agent to create (outbound, inbound, or widget).
Agent Name The name of the agent to be created.
Prompt The prompt the agent will use during conversations.
Greeting Message Greeting message the agent will use at the start of the conversation.
LLM Language model identifier to be used by the agent (e.g., gpt-5.1 or gpt-5).
Language Language code the agent should use (e.g., en, de).
Voice ID Identifier of the voice the agent will use.
Additional Agent Fields Additional optional fields to customize the agent, including raw JSON configuration, consent message, description, webhook URLs, recording and consent settings, max call duration, and phone number.

Output

JSON

  • type - The type of the created agent (outbound, inbound, or widget).
  • name - The name of the created agent.
  • agent
    • prompt - The prompt used by the agent during conversations.
    • greeting_message - The greeting message used by the agent at the start of conversations.
    • llm - The language model identifier used by the agent.
    • language - The language code used by the agent.
    • voice_id - The voice identifier used by the agent.
    • consent_message - Consent message spoken by the agent, if provided.
    • is_consent_enabled - Boolean indicating if consent to record calls is enabled.
  • description - Description of the agent, if provided.
  • phone_number - Phone number associated with the agent, if provided.
  • external_webhook_url - Webhook URL to receive post-call data, if provided.
  • inbound_call_webhook_url - Webhook URL to receive inbound call events, if provided.
  • is_recording - Boolean indicating if call recording is enabled.
  • max_duration
    • duration_seconds - Maximum call duration in seconds, if set.
    • is_enabled - Boolean indicating if max call duration limit is enabled.

Dependencies

  • Synthflow API authentication credential

Troubleshooting

  • Invalid JSON in the 'Agent JSON' field will cause an error; ensure the JSON is valid and represents an object.
  • Missing required fields such as Agent Name, Prompt, Greeting Message, LLM, Language, or Voice ID will cause the operation to fail.
  • Network or authentication errors when calling the Synthflow API can occur; verify API credentials and network connectivity.

Links

Discussion