Actions: AI

GPT chat completion

This submits the prompt to the GPT Chat Completion endpoint. It includes also, the previous messages and responses in the conversation. These can be accessed independently using the ‘GPTConversations’ array functions. The documentation for this endpoint can be found here: https://platform.openai.com/docs/api-reference/chat/create.

Argument Description
ResponseOutput Where the resulting response text should be be placed.
ResponseSchema JSON Schema for structured response. To learn more, and to use our helper bot, visit https://platform.openai.com/docs/guides/structured-outputs/
ConversationName The name that will be used to track this conversation. This name gives you the ability to access the conversation’s responses directly using the ‘GPTConversations’ variable array.
Model The model that should be used for the request. If you use ‘chatgpt-4o-latest’ your code will always use the latest 4o model. Visit https://platform.openai.com/docs/models for more info.
BasePrompt The prompt which defines the desired behavior of the AI. This is where you tell the AI what role to play in this conversation. Describe it in as much detail as you can. For example: You are a world class designer, and you will help me design an app based on proper design rules and conventions.
RequestPrompt Where the user’s next prompt should come from

GPT delete file

Delete an uploaded file from the OpenAI File API.

Argument Description
FileID Variable containing the ID of the file to delete.
DeleteResult Variable to store API response JSON as array.

GPT file info

Retrieve metadata for a specific OpenAI file by its file ID.

Argument Description
FileID Select a variable that contains the OpenAI file ID.
FileInfo An array containing the information of the file.

GPT list files

List files accessible to the configured OpenAI account.

Argument Description
Files Array of the file objects present in your ORGs File API

GPT response

Unified Responses API with conversation state, optional file attachments, and optional inline image input.

Argument Description
DetailMode Optional vision detail mode for inline images: auto, low, high.
ImageData Include images in the request. Supports multiple files. Each line can be a URL to a remote file, a BaseEncodedImage, or a FileID from the Open AI Files API
ResponseSchema Optional JSON schema for structured output.
Model The model to use for the response.
BasePrompt Define how the AI should behave and respond in this conversation.
ResponseOutput Where to place the response output. The full JSON response is available in the GPTConversations array.
ConversationName Name of the conversation thread in GPTConversations.
RequestPrompt Next prompt from the user.
VectorStoreIDs Optional variable holding a comma delimited list of Vector Store IDs from gptFileUpload action
ReasoningEffort When using reasoning models like GPT-5 or o3, you can specify the amount of reasoning effort the model puts in.

GPT upload file

Uploads a file you already have in a variable to the OpenAI Files API. Returns file_id and the full JSON response.

Argument Description
VectorStoreID Variable to receive the created vector_store id.
ResponseOutput Where to store the full response (Variable or Element).
FileName Filename to send in multipart Content-Disposition (e.g., report.pdf).
Purpose OpenAI file purpose (e.g., assistants, batch, fine-tune). Choose Variable or Input.
LifeTime
FileData Variable that contains the raw file data to upload.
FileID Variable to receive the returned file_id string.
CreateVectorStore Whether to create a vector embedding for your file. This is required to use documents with GPT Response action.

GPT vision

This sends an image to the OpenAI Vision API, returning the GPTs response to your image based on the prompt. The call is synchronous.

Argument Description
ResponseSchema JSON Schema for structured response. To learn more, and to use our helper bot, visit https://platform.openai.com/docs/guides/structured-outputs/
ResponseOutput Where the response text will be placed.
DetailMode The input detail mode. Low = 512x512, High = 768x2048 or 2048x768. Input images are resized to those dimensions by OpenAI. It is best to aim to send the appropriate image dimensions. Limit per image is 12 MB.
ImageData The binary data of the image. Your image will be resized based on the detail level you selected. When this variable is a numerical array, we use will use ‘ImageData’ as the key that contains the ImageData
ResponseModel The GPT model to use to generate the text response.
TopicName The name that will be used to track calls. This name gives you the ability to access the returned data directly using the ‘GPTVision’ array variable.
RequestPrompt The prompt for the request. ie: What is in the image?
BasePrompt The prompt which defines the desired behavior of the AI. This is where you tell the AI what role to play in this conversation. Describe it in as much detail as you can. For example: You are a world class designer, and you will help me design an app based on proper design rules and conventions.

clear GPT conversation

Clear the content of a GPT Conversation.

Argument Description
ConversationName Select the GPT conversation to be cleared.

convert text to speech

This converts text to speech using OPENAIs Text To Speech API. Documentation for this endpoint can be found here: https://platform.openai.com/docs/api-reference/audio/createSpeech.

Argument Description
Response Where the resulting audio data be be placed
Text The text to be turned into audio.
Model The model that should be used for the request.
Voice The OPENAI TTS Voice to be used.
Speed The speed of the speech.

generate image

Generates an image using OpenAI API and places the image in a variable or image. The documentation for this endpoint can be found here: https://platform.openai.com/docs/api-reference/images.

Argument Description
ResponseFormat Return Image BINARY Data or a URL.
Resolution The resolution of the resulting image
RequestPrompt The prompt for the image generation request. This is where you describe the image you wish to generate in as much detail as you can.
Style Generation style to use.
ConversationName The name that will be used to track this conversation. This name gives you the ability to access the conversation’s array directly using the GPTConversations array.
Quality The quality setting to use for the request. ‘standard’ produced lower quality results, and ‘hd’ produces the highest quality images the DALE-3 model is capable of.
ResponseOutput Where the URL or image data should go.

This chapter was last updated on Fri 24 Oct 2025 16:52:19 BST