Skip to main content

LLM Steps

Steps in the LLM category.

Available Steps​

AZURE_OPENAI_COMPLETION​

Chat/completion step that calls OpenAI (or Hub/Azure variants). Supports messages, images, audio, attachments, tools (functions), structured outputs (json_schema/json_object/list), reasoning and audio output. Many step-level configuration options are supported via cfg.*.

Aliases: SERVICE_AZURE_OPENAI_COMPLETION

Version: 1.0.0

Example:


- step: AZURE_OPENAI_COMPLETION
  behavior: "You are a helpful assistant."
  instruction: "Write a short summary of the benefits of unit testing. Answer in JSON format."
  options:
    temperature: 0.2
  response_format: "json_object"

AZURE_OPENAI_IMAGE_GENERATION​

ImageGeneration using Azure OpenAI.

Aliases: SERVICE_AZURE_OPENAI_IMAGE_GENERATION

Version: 1.0.0

Example:


- step: SET
  text: "A futuristic skyline at sunset, cinematic lighting"
- step: AZURE_OPENAI_IMAGE_GENERATION
  n: 2
  size: 1024x1024

AZURE_OPENAI_SPEECH​

Azure OpenAI speech step alias.

Aliases: SERVICE_AZURE_OPENAI_SPEECH

Version: 1.0.6

Example:


- step: AZURE_OPENAI_SPEECH
  model: "tts-1"
  voice: "alloy"
  response_format: "mp3"

AZURE_OPENAI_TRANSCRIPTION​

Azure OpenAI transcription alias (delegates to the OpenAI transcription implementation).

Aliases: SERVICE_AZURE_OPENAI_TRANSCRIPTION

Version: 1.0.0

Example:


- step: AZURE_OPENAI_TRANSCRIPTION
  model: "whisper-1"

HUB_COMPLETION​

Chat/completion step that calls OpenAI (or Hub/Azure variants). Supports messages, images, audio, attachments, tools (functions), structured outputs (json_schema/json_object/list), reasoning and audio output. Many step-level configuration options are supported via cfg.*.

Version: 1.0.0

Example:


- step: HUB_COMPLETION
  behavior: "You are a helpful assistant."
  instruction: "Write a short summary of the benefits of unit testing. Answer in JSON format."
  options:
    temperature: 0.2
  response_format: "json_object"

HUB_CONTENT_EXTRACTION​

Extract textual content from a URL using the ProActions-Hub extraction tool. The URL can be provided via the default text input or the in/url configuration.

Version: 1.0.2

Example:


- step: SET
  text: "https://example.com/article"
- step: HUB_CONTENT_EXTRACTION

HUB_IMAGE_GENERATION​

ImageGeneration using ProActions Hub.

Version: 1.0.0

Example:


- step: SET
  text: "A futuristic skyline at sunset, cinematic lighting"
- step: HUB_IMAGE_GENERATION
  n: 2
  size: 1024x1024

HUB_SPEECH​

ProActionsHub-compatible speech generation step. Inherits behavior from OPENAI_SPEECH.

Version: 1.0.6

Example:


- step: HUB_SPEECH
  model: "tts-1"
  voice: "alloy"
  response_format: "mp3"

HUB_TRANSCRIPTION​

Hub-compatible transcription step (delegates to hub service). Writes transcription text and optional segments list.

Aliases: SERVICE_OPENAI_TRANSCRIPTION

Version: 1.0.0

Example:


- step: HUB_TRANSCRIPTION
  model: "whisper-1"

OPENAI_ASSISTANT​

Create or reuse an OpenAI assistant. If cfg.assistantId is provided the assistant will be retrieved; otherwise a new assistant is created. The created or retrieved assistant is written to the flow context.

Version: 1.0.9

Example:


- step: OPENAI_ASSISTANT
  assistantId: asst_Ja0fm7Rova3QJGbKrAKrt7HS

OPENAI_COMPLETION​

Chat/completion step that calls OpenAI (or Hub/Azure variants). Supports messages, images, audio, attachments, tools (functions), structured outputs (json_schema/json_object/list), reasoning and audio output. Many step-level configuration options are supported via cfg.*.

Aliases: SERVICE_OPENAI_COMPLETION

Version: 1.0.0

Example:


- step: OPENAI_COMPLETION
  behavior: "You are a helpful assistant."
  instruction: "Write a short summary of the benefits of unit testing. Answer in JSON format."
  options:
    temperature: 0.2
  response_format: "json_object"

OPENAI_DELETE_THREAD​

Delete a thread using its id. Provide the thread object or id via inputs.

Version: 1.0.9

Example:


- step: OPENAI_DELETE_THREAD
  thread: "{{threadObject}}"

OPENAI_IMAGE_GENERATION​

Generate images using OpenAI image models (DALL·E and GPT-Image-1). Supports model, size, quality, count (n), response_format and other parameters. The generated images are stored in the configured output key or default imageList.

Aliases: SERVICE_OPENAI_IMAGE_GENERATION

Version: 1.0.0

Example:


- step: SET
  text: "A futuristic skyline at sunset, cinematic lighting"
- step: OPENAI_IMAGE_GENERATION
  n: 2
  size: 1024x1024

OPENAI_SPEECH​

Generate speech audio using the OpenAI Speech API (or corresponding Hub/Azure variants). Reads text from the default text input and returns an audio blob.

Aliases: SERVICE_OPENAI_SPEECH

Version: 1.0.6

Example:


- step: OPENAI_SPEECH
  model: "tts-1"
  voice: "alloy"
  response_format: "mp3"

OPENAI_THREAD​

Create or reuse an OpenAI thread. If cfg.reuse and cfg.storeIn are provided the thread id may be reused. The created thread is written to the flow context.

Version: 1.0.9

Example:


- step: OPENAI_THREAD
  reuse: true
  storeIn: "session"

OPENAI_THREAD_FILES​

Create a vector store for thread file tools and optionally upload files. Writes created store to flow context and optional uploaded file list to an optional output.

Version: 1.0.9

Example:


- step: OPENAI_THREAD_FILES
  replaceStores: true
  expires_after: 1
  name: "my-store"

OPENAI_THREAD_MESSAGE​

Send a message to a thread/assistant. The assistant and thread inputs are required. The message can be provided via cfg.instruction or via default text input. The assistant response text is stored in the default text output.

Version: 1.0.9

Example:


- step: OPENAI_ASSISTANT # load assistant
  assistantId: asst_Ja0fm2Rova2QJGbKrAKrtXXX
- step: OPENAI_THREAD # load an existing thread
  reuse: true
  storeIn: session
- step: OPENAI_THREAD_MESSAGE
  instruction: |
    Create a list of the key takeaways from the documents provided.

OPENAI_TRANSCRIPTION​

Transcribe audio using OpenAI (or Azure/Hub variants). Reads an audio file/blob from inputs and writes the transcription text and optional segment list.

Aliases: SERVICE_OPENAI_TRANSCRIPTION

Version: 1.0.0

Example:


- step: OPENAI_TRANSCRIPTION
  model: "whisper-1"

STABILITY_AI_OUTPAINT​

Outpaint an image region using StabilityAI. Supports coordinates (left,right,up,down) to define the outpainting area.

Aliases: SERVICE_STABILITY_AI_OUTPAINT

Version: 1.0.4

Example:


- step: STABILITY_AI_OUTPAINT
  left: 10
  right: 10
  up: 20
  down: 20

STABILITY_AI_SEARCH_AND_RECOLOR​

Search and recolor elements in an image using StabilityAI. Provide select_prompt and prompt to find and recolor elements.

Aliases: SERVICE_STABILITY_AI_SEARCH_AND_RECOLOR

Version: 1.0.4

Example:


- step: STABILITY_AI_SEARCH_AND_RECOLOR
  select_prompt: "select all trees"
  prompt: "make them autumn color"

STABILITY_AI_SEARCH_AND_REPLACE​

Search and replace regions or objects within an image using StabilityAI. Provide a prompt and a search_prompt to locate and replace visual elements.

Aliases: SERVICE_STABILITY_AI_SEARCH_AND_REPLACE

Version: 1.0.4

Example:


- step: STABILITY_AI_SEARCH_AND_REPLACE
  search_prompt: "replace red car"
  prompt: "a blue car instead"

STABILITY_AI_UPSCALE​

Upscale an image using StabilityAI upscale endpoint. Uploads a provided image blob and returns the upscaled image blob.

Aliases: SERVICE_STABILITY_AI_UPSCALE

Version: 1.0.4

Example:


- step: STABILITY_AI_UPSCALE
  output_format: "jpeg"