Send prompts to OpenAI's ChatGPT API and use the response dynamically in your flow.
Send prompts to OpenAI's ChatGPT API and use the response dynamically in your flow.
Category: AI
Tutorial: Start interactive node tutorial
| Label | Property | Type | Default | Description |
|---|---|---|---|---|
| System prompt | systemPrompt | textarea | "" | The system prompt |
| User prompt | userPrompt | textarea | "" | The user prompt with optional handlebar variables |
| Model | model | select | "gpt-4.1-mini" | Specify the model to use |
| Collect and send history | manageHistory | checkbox | true | If checked the node will manage the history on his own |
| Response Format | responseFormat | textarea | "" | The openAi response_format definition as JSON text |
| On back in context | onReentry | select | "reset" | What should happen when the node re-enters the context? |
| On prompt change | onChange | select | "reset" | What should happen if the user prompt changes? |
| State | Initial Value | Description |
|---|---|---|
| done | false | True if request has finished |
| thinking | false | True if request is running |
| status | 0 | The response HTTP status code |
| answer | null | The LLM answer string |
| error | false | True if an error occurred |
| lastUserPrompt | "" | Contains the last user prompt |
| history | Contains the history if managed by the node | |
| lastUsage | {} | Contains the token usage statistics for the last request |
| using | false | Is the model using a tool right now? |
reset() - Reset to initial statesend() - Send requestserverSideSend() - Send the request to the server sideused() - Called from a tool node after successful use of the tool.