Skip to main content

History for GPT Actions in Stack

What's New?​

We're thrilled to introduce a powerful enhancement to our GPT actions: the ability to retain and utilize history. Now, GPT actions in Stack can generate more relevant and personalized responses by considering past interactions.

Types of History​

There are five types of history that can be stored:

  • This Sub Account: Keeps track of all GPT conversations across all workflows within the sub-account where the history type is set to "This Sub Account".
  • This Workflow: Remembers all GPT conversations within a specific workflow where the history type is "This Workflow".
  • Per Execution: Stores GPT conversations within a workflow for a single execution when the history type is "Per Execution".
  • This Step: Retains GPT conversations for a specific action across multiple executions when the history type is "This Step".
  • Custom: Allows you to create a custom history type that can be used within the same workflow or across multiple workflows.

How to Use​

  1. Enable the option in Labs.
  2. Select the GPT Powered By OpenAI action.
  3. Turn on Enable history.
  4. Choose the History for option from the dropdown menu.
  5. Add any necessary system instructions.

Advanced Options​

  • System Instructions: You can add specific instructions to guide the GPT action for more precise and desired outputs. These rules will be applied to all GPT actions.
  • Exclude Instructions from History: Use this toggle if you want to prevent system instructions from being stored in the history.
  • Exclude Responses from History: Turn this on if you don't want the responses from this action to be stored. This can be useful for responses that are sentiment-based or include analytics.

Points to Note​

  • When history is enabled, GPT-4 models are selected by default.
  • History functionality is available only for the "Custom" action type.
  • History is maintained independently for each contact.