How to Use GPT History in Sparkforge Workflow Actions

By Sparkforge Support Team ·Last reviewed May 3, 2026

GPT History lets the GPT Powered by OpenAI action remember prior prompts and responses across workflows, executions, or steps so future calls produce more contextual replies.

What this does

The GPT Powered by OpenAI action can store conversation history so future calls reference earlier prompts and outputs. Result: more contextual, more personalized responses for that contact.

History scopes

  • This Sub Account — Remembers GPT calls across every workflow in the account.
  • This Workflow — Remembers GPT calls inside the same workflow.
  • Per Execution — Remembers only within a single workflow execution.
  • This Step — Remembers calls for the same action across executions.
  • Custom — Define a custom history bucket reusable across workflows.

How to enable

  1. Turn on the option in Labs.
  2. Add the GPT Powered by OpenAI action.
  3. Toggle Enable History on.
  4. Pick a History for scope from the dropdown.
  5. Add system instructions if needed.

Advanced options

  • System Instructions — Rules every GPT action obeys, useful for tone, persona, or format constraints.
  • Exclude Instructions from History — Keep system rules out of the saved history bucket.
  • Exclude Responses from History — Don't save outputs to history. Useful when responses are sentiment scores or analytics, not conversation.

Things to know

  • Enabling history defaults the model to GPT-4.
  • History only works with the Custom action type.
  • History is independent for each contact.

Example: recommend products to existing customers

Trigger: Order Placed

  1. GPT Powered by OpenAI — multiple actions sharing history. Use them to draft an email and analyze sentiment of the reply.
  2. Email — Send the GPT-generated message to the customer.
  3. Wait — For the customer reply.
  4. If/Else — Branch on sentiment. Positive replies continue the conversation; negative replies exit the workflow.

Frequently Asked Questions

Which scope should I pick?
Use Per Execution for one-shot conversations, This Workflow when you want continuity inside a flow, and This Sub Account when GPT actions across the entire account should share memory for the same contact.
Can I keep system rules out of saved history?
Yes. Toggle Exclude Instructions from History on. The rules still apply at runtime, they just don't bloat the saved transcript.
Why would I exclude responses from history?
When the GPT call returns analytics — sentiment, classification, scoring — you don't want those answers polluting the conversation memory used by other actions.
Does history work across contacts?
No. Each contact has independent history. One contact's prior conversation never bleeds into another contact's GPT calls.

Was this article helpful?