Orq Agent icon

Orq Agent

Invoke an Orq AI Agent

Orq Agent

AI-generated

Overview

This node invokes an Orq AI Agent to send a message and receive a response. It is useful for integrating AI-driven conversational agents into workflows, allowing users to send messages to the agent, manage conversation threading, and handle agent responses including refusals and incomplete statuses. Practical examples include customer support automation, AI-driven chatbots, and interactive virtual assistants.

Use Case Examples

  1. Sending a message to an Orq AI Agent to get a response based on user input.
  2. Continuing a conversation by specifying a previous response ID or conversation ID to maintain context.
  3. Using memory entities to allow the agent to recall facts across multiple calls.

Properties

Name Meaning
Agent Select the Orq AI Agent key to send the message to.
Message The message text to send to the agent.
Additional Fields Optional parameters including previous response ID, conversation ID, memory entity ID, whether to store the response, template variables, and metadata tags.
Timeout (Seconds) HTTP timeout for the agent call, defaulting to 600 seconds (10 minutes) to accommodate long-running agent executions.

Output

JSON

  • responseId - Unique identifier of the agent's response.
  • agentKey - The agent key used for the request.
  • status - Status of the response, e.g., completed, incomplete, failed.
  • success - Boolean indicating if the response was successfully completed.
  • response - The textual response from the agent, concatenated from output messages.
  • raw - The full raw response object from the Orq API.
  • usage - Optional usage statistics related to the agent call.
  • refusals - List of refusal reasons if the agent refused to answer certain parts.
  • incomplete - Boolean indicating if the response is incomplete.
  • incompleteReason - Reason for the response being incomplete, if applicable.

Dependencies

  • Requires an Orq API key credential for authentication.

Troubleshooting

  • Ensure the agent key and message are valid and not empty to avoid validation errors.
  • Check that only one of Previous Response ID or Conversation ID is provided to avoid exclusivity validation errors.
  • If the response status is 'failed', inspect the error message for details and retry or adjust inputs accordingly.
  • If the response status is 'requires_action', additional user intervention may be needed as indicated by the API.
  • Timeouts can occur if the agent takes longer than the specified timeout; increase the timeout if necessary.
  • If the node throws a request failed error, verify network connectivity and API credentials.

Links

Discussion