OpenAI Node Enhancements

Proposed Enhancements

  1. Dynamic Message Retrieval from Memory Node or Variables:
    Implement functionality to dynamically fetch messages from a specified memory node or variables, in addition to the current method of using fixed messages. This would allow for more flexible and context-aware interactions within workflows.

  2. Message Templates for System Prompts:
    Introduce message templates specifically designed for system prompts, enabling users to quickly and efficiently generate standardized system messages that can be customized and reused across various workflows.

  3. Tool Calling Functions Integration:
    Enable the OpenAI node to call tool functions directly within the workflow, facilitating a tighter integration of OpenAI’s capabilities with other operational aspects of the n8n environment, such as data processing or analysis tasks.

  4. Support for Image and Audio Content Interpretation:
    Allow the OpenAI node to accept and interpret image and audio content directly. Images would be processed as-is (GPT-4-Vision), while audio content would be automatically sent to the Whisper API for transcription before being used as a message in the conversation. This feature would greatly expand the types of content that can be handled within conversational AI workflows.

  5. Display of Token Usage in the Editor:
    Implement a feature to display the number of tokens used by the OpenAI node in the n8n editor. This would help users manage their API usage more effectively, providing insights into the cost implications of their configurations.