OpenAI

OpenAI API

This node signs in with a credential: OpenAI API Key. See Connect a credential.

Actions

Steps this node can perform. Add one from Add a step, then pick the action.

Text Message

Chat with AI

Settings

Field	Type	Required	Description
Model	Choice	Yes	The model to use Options: GPT-5.2, GPT-5.1, GPT-5, GPT-5 Mini, GPT-5 Nano, GPT-4.1, GPT-4.1 Mini, GPT-4.1 Nano, GPT-5.5 Default: gpt-5.5
Messages	List	Yes	The messages to send to the model
Temperature	Number		The temperature to use Default: 0.5
Top P	Number		The top P to use Default: 0.5
Frequency Penalty	Number		The frequency penalty to use Default: 0
Presence Penalty	Number		The presence penalty to use Default: 0
Max Tokens	Number		The maximum number of tokens to generate Default: 256
Stop	Text		The stop sequence to use
Reasoning Effort	Choice		The reasoning effort to use Options: Low, Medium, High Default: low
Verbosity	Choice		The verbosity to use Options: Low, Medium, High Default: low

Output

Later steps use this data by picking this step in the $ picker. For example, model_used is $.<step ID>.output.model_used.

{
  "model_used": "string",
  "response": {
    "choices": [
      {
        "finish_reason": "string",
        "index": 0,
        "message": {}
      }
    ],
    "created": 0,
    "id": "string",
    "model": "string",
    "object": "chat.completion",
    "usage": {
      "completion_tokens": 0,
      "prompt_tokens": 0,
      "total_tokens": 0
    }
  },
  "success": false,
  "tokens_used": 0
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
model_used	Text	The model that produced the chat completion
response	Object	The full chat completion response returned by OpenAI
response.choices[]	List	A list of chat completion choices
response.choices[].finish_reason	Text	The reason the model stopped generating tokens
response.choices[].index	Number	The index of the choice in the list
response.choices[].message	Object	The message generated by the model
response.choices[].message.content	Text	The contents of the message
response.choices[].message.role	Text	The role of the message author
response.created	Number	The Unix timestamp (in seconds) of when the chat completion was created
response.id	Text	A unique identifier for the chat completion
response.model	Text	The model used for the chat completion
response.object	Text	The object type, which is always 'chat.completion'
response.usage	Object	Usage statistics for the completion request
response.usage.completion_tokens	Number	Number of tokens in the generated completion
response.usage.prompt_tokens	Number	Number of tokens in the prompt
response.usage.total_tokens	Number	Total number of tokens used in the request
success	Yes / No	Whether the chat completion request succeeded
tokens_used	Number	Total number of tokens used in the request

Deep Research

Create a new response using OpenAI's Deep Research API for stateful conversations

Settings

Field	Type	Required	Description
Model	Choice	Yes	The model to use for the response Options: GPT-5.6 Sol Default: gpt-5.6-sol
Input	Text	Yes	Input text or messages for the response
Tools	List	Yes	Built-in tools to use (web_search)
Store Response	Yes / No		Whether to store the response in the database Default: false
Max Tool Calls	Number		Maximum number of tool calls Default: 10
Previous Response ID	Text		ID of the previous response to use as context
Metadata	Object		Metadata for the response
Reasoning Effort	Choice		Reasoning effort to use Options: Medium, High Default: medium
Execution Mode	Choice		How the action handles the deep research job. 'Background Request' submits the job and returns immediately with a response ID — the flow continues without waiting for results. 'Wait Until Finish' submits the job and polls until it reaches a terminal state (completed, failed, or cancelled) before passing control to the next node. Options: Background Request, Wait Until Finish Default: background_request
Polling Interval (seconds)	Number		Time in seconds between status checks when Execution Mode is 'Wait Until Finish'. Default is 60 seconds. OpenAI recommends at least 2 seconds between polls. Default: 60

Output

Later steps use this data by picking this step in the $ picker. For example, conversation_id is $.<step ID>.output.conversation_id.

{
  "conversation_id": "string",
  "created_at": 0,
  "error": {
    "message": "string",
    "type": "string"
  },
  "id": "string",
  "metadata": {},
  "model": "string",
  "object": "response",
  "output": [
    {
      "content": "string",
      "role": "string"
    }
  ],
  "status": "in_progress",
  "usage": {
    "completion_tokens": 0,
    "prompt_tokens": 0,
    "total_tokens": 0
  }
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
conversation_id	Text	ID of the conversation this response belongs to
created_at	Number	The Unix timestamp (in seconds) of when the response was created
error	Object	Error information if the response failed
error.message	Text	Error message
error.type	Text	Error type
id	Text	A unique identifier for the response
metadata	Object	Additional metadata for the response
model	Text	The model used for the response
object	Text	The object type, which is always 'response'
output[]	List	Array of output messages or content
output[].content	Text	The content of the message
output[].role	Text	The role of the message author
status	Text	The current status of the response
usage	Object	Usage statistics for the response
usage.completion_tokens	Number	Number of tokens in the generated completion
usage.prompt_tokens	Number	Number of tokens in the prompt
usage.total_tokens	Number	Total number of tokens used

Create Response

Create a new response using OpenAI's Responses API for stateful conversations

Settings

Field	Type	Required	Description
Model	Choice	Yes	The model to use for the response Options: GPT-5.2, GPT-5.1, GPT-5, GPT-5 Mini, GPT-5 Nano, GPT-4.1, GPT-4.1 Mini, GPT-4.1 Nano, GPT-5.5, GPT-5.5 Pro Default: gpt-5.5
Input	Text	Yes	Input text or messages for the response
Tools	List		Built-in tools to use (web_search)
Stream Response	Yes / No		Enable streaming responses Default: false
Temperature	Number		Controls randomness (0.0 to 2.0) Default: 1
Max Output Tokens	Number		Maximum tokens to generate Default: 1000
Instructions	Text		Instructions for the response
Store Response	Yes / No		Whether to store the response in the database Default: false
Max Tool Calls	Number		Maximum number of tool calls Default: 10
Previous Response ID	Text		ID of the previous response to use as context
Metadata	Object		Metadata for the response
Reasoning Effort	Choice		Reasoning effort to use Options: Low, Medium, High Default: low
Background Processing	Yes / No		Whether to process the response in the background. When true, the API returns immediately with a response ID that can be used to retrieve the result later. Default: false
Response Format	Object		Specifies the format of the model's output. Use for structured outputs with JSON Schema.

Output

Later steps use this data by picking this step in the $ picker. For example, conversation_id is $.<step ID>.output.conversation_id.

{
  "conversation_id": "string",
  "created_at": 0,
  "error": {
    "message": "string",
    "type": "string"
  },
  "id": "string",
  "metadata": {},
  "model": "string",
  "object": "response",
  "output": [
    {
      "content": "string",
      "role": "string"
    }
  ],
  "status": "in_progress",
  "usage": {
    "completion_tokens": 0,
    "prompt_tokens": 0,
    "total_tokens": 0
  }
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
conversation_id	Text	ID of the conversation this response belongs to
created_at	Number	The Unix timestamp (in seconds) of when the response was created
error	Object	Error information if the response failed
error.message	Text	Error message
error.type	Text	Error type
id	Text	A unique identifier for the response
metadata	Object	Additional metadata for the response
model	Text	The model used for the response
object	Text	The object type, which is always 'response'
output[]	List	Array of output messages or content
output[].content	Text	The content of the message
output[].role	Text	The role of the message author
status	Text	The current status of the response
usage	Object	Usage statistics for the response
usage.completion_tokens	Number	Number of tokens in the generated completion
usage.prompt_tokens	Number	Number of tokens in the prompt
usage.total_tokens	Number	Total number of tokens used

Retrieve Response

Retrieve the status and results of a response from OpenAI's Responses API

Settings

Field	Type	Required	Description
Response ID	Text	Yes	ID of the response to retrieve

Output

Later steps use this data by picking this step in the $ picker. For example, conversation_id is $.<step ID>.output.conversation_id.

{
  "conversation_id": "string",
  "created_at": 0,
  "error": {
    "message": "string",
    "type": "string"
  },
  "id": "string",
  "metadata": {},
  "model": "string",
  "object": "response",
  "output": [
    {
      "content": "string",
      "role": "string"
    }
  ],
  "status": "in_progress",
  "usage": {
    "completion_tokens": 0,
    "prompt_tokens": 0,
    "total_tokens": 0
  }
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
conversation_id	Text	ID of the conversation this response belongs to
created_at	Number	The Unix timestamp (in seconds) of when the response was created
error	Object	Error information if the response failed
error.message	Text	Error message
error.type	Text	Error type
id	Text	A unique identifier for the response
metadata	Object	Additional metadata for the response
model	Text	The model used for the response
object	Text	The object type, which is always 'response'
output[]	List	Array of output messages or content
output[].content	Text	The content of the message
output[].role	Text	The role of the message author
status	Text	The current status of the response
usage	Object	Usage statistics for the response
usage.completion_tokens	Number	Number of tokens in the generated completion
usage.prompt_tokens	Number	Number of tokens in the prompt
usage.total_tokens	Number	Total number of tokens used

Generate Image

Generate an image from a text prompt using OpenAI's GPT-image models

Settings

Field	Type	Required	Description
Prompt	Text	Yes	Text description of the desired image(s). Max 32000 characters for GPT-image models.
Model	Choice	Yes	The model to use for image generation. DALL-E models are deprecated and will stop being supported on 05/12/2026. Options: GPT Image 1.5 (Recommended), GPT Image 1, GPT Image 1 Mini Default: gpt-image-1.5
Quality	Choice		Image quality level. Lower quality images generate faster. Only supported for GPT-image models. Options: High, Medium, Low Default: medium
Resolution	Choice		Resolution of the generated image. Square images are faster to generate. Options: 1024x1024 (Square), 1024x1536 (Portrait), 1536x1024 (Landscape) Default: 1024x1024
Background	Choice		Background transparency. Only supported for GPT-image models. Requires png or webp output format for transparency. Options: Auto, Opaque, Transparent Default: auto
Number of Images	Number		Number of images to generate (1-10). DALL-E 3 only supports n=1. Default: 1
Moderation	Choice		Content moderation sensitivity level. Only supported for GPT-image models. Options: Auto (Standard filtering), Low (Less restrictive) Default: auto
Output File	File	Yes	The configuration for the output file. The file extension will be determined by the output_type field
Output Type	Choice		The image file format for the output Options: PNG, WebP, JPEG Default: png

Output

Later steps use this data by picking this step in the $ picker. For example, created is $.<step ID>.output.created.

{
  "created": 0,
  "data": [
    {
      "fileId": "string",
      "fileName": "string",
      "fileSize": 0
    }
  ],
  "model": "string",
  "prompt": "string"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
created	Number	When OpenAI generated the images, in Unix seconds
data[]	List	Generated images, saved to file storage
data[].fileId	Text	ID of the saved image file
data[].fileName	Text	Name of the saved image file
data[].fileSize	Number	Image size in bytes
model	Text	Image model used
prompt	Text	Prompt used to generate the images

Classify Text for Violations

Identify and flag content that might be harmful using OpenAI's moderation API

Settings

Field	Type	Required	Description
Text Input	Text	Yes	Enter text to classify if it violates the moderation policy
Options	Object		Additional options for text moderation

Output

Later steps use this data by picking this step in the $ picker. For example, flagged is $.<step ID>.output.flagged.

{
  "flagged": false,
  "model_used": "string",
  "response": {
    "id": "string",
    "model": "string",
    "results": [
      {
        "categories": "hate",
        "category_scores": {},
        "flagged": false
      }
    ]
  },
  "success": false
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
flagged	Yes / No	Whether the first input was flagged; absent when OpenAI returned no results
model_used	Text	Moderation model OpenAI used
response	Object	The full OpenAI moderation response
response.id	Text	Moderation request ID
response.model	Text	Moderation model
response.results[]	List	One result per input
response.results[].categories	Object	Per-category verdicts, e.g. hate, harassment, violence
response.results[].category_scores	Object	Per-category scores from 0 to 1
response.results[].flagged	Yes / No	Whether the input violates the usage policies
success	Yes / No	True when the moderation request succeeded

Upload File

Upload files to OpenAI for use with Assistants, Batch processing, Fine-tuning, and Vision models

Settings

Field	Type	Required	Description
Purpose	Choice	Yes	Purpose of the file upload Options: Assistants (512 MB max), Vision (20 MB max), Batch (100 MB max), Fine-tune (1 GB max) Default: batch
File Content Type	Choice	Yes	Type of file content - 'reference' for FSS storage key or 'raw' for base64 data Options: Raw Content. With big data please consider using reference, Reference to a file in Storage Default: raw
File Content	Text	Yes	Base64-encoded file content (if file_content_type is 'raw') or Storage Key (if file_content_type is 'reference')
Filename	Text	Yes	Original filename with extension (e.g., document.pdf, image.png)
Content Type	Text		MIME type of the file (optional, auto-detected if not provided)

Output

Later steps use this data by picking this step in the $ picker. For example, bytes is $.<step ID>.output.bytes.

{
  "bytes": 0,
  "file_id": "string",
  "filename": "string",
  "purpose": "assistants"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
bytes	Number	Size of the uploaded file in bytes
file_id	Text	Unique identifier for the uploaded file
filename	Text	Name of the uploaded file
purpose	Text	Purpose of the uploaded file

Transcribe Audio

Transcribe an attached or same-run reference audio file using the direct OpenAI API

Settings

Field	Type	Required	Description
Model	Choice	Yes	OpenAI transcription model Options: GPT Transcribe, GPT-4o Transcribe, GPT-4o Mini Transcribe, Whisper 1, GPT-4o Transcribe Diarize Default: whisper-1
Audio File	File	Yes	Audio file to transcribe (FLAC, MP3, MP4, MPEG, MPGA, M4A, OGG, WAV, or WEBM; max 25 MiB) Options: Reference, Attach new
Response Format	Choice		Format of the transcription response. Use 'diarized_json' for speaker separation with gpt-4o-transcribe-diarize. Options: JSON, Diarized JSON Default: json
Prompt	Text		Optional text to guide the model's style or vocabulary (e.g. proper nouns). Note: not supported for diarization models.
Chunking Strategy	Choice		Chunking strategy for audio. Set to 'auto' for inputs longer than 30 seconds with gpt-4o-transcribe-diarize. Options: Auto

Output

Later steps use this data by picking this step in the $ picker. For example, model_used is $.<step ID>.output.model_used.

{
  "model_used": "string",
  "segments": [
    {
      "end": 0,
      "id": "string",
      "speaker": "Speaker 1",
      "start": 0,
      "text": "string"
    }
  ],
  "text": "string"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
model_used	Text	OpenAI transcription model used
segments[]	List	Segments with timestamps and speaker labels when using diarized_json or verbose_json
segments[].end	Number	End timestamp in seconds
segments[].id	Any	Segment ID
segments[].speaker	Text	Speaker label (e.g. Speaker 1, Speaker 2)
segments[].start	Number	Start timestamp in seconds
segments[].text	Text	Transcribed segment text
text	Text	Transcribed audio text

Create Batch Job

Create batch jobs for asynchronous processing of large volumes of requests at 50% reduced cost. Requires a file uploaded with purpose 'batch'. Completion window is automatically set to 24h. Automatically polls batch status every 1 minute until it reaches 'in_progress' or 'failed' state.

Settings

Field	Type	Required	Description
Input File ID	Text	Yes	The ID of an uploaded file that contains requests for the new batch. File must be in JSONL format, uploaded with purpose 'batch', contain up to 50,000 requests, and be up to 200 MB in size.
Endpoint	Choice	Yes	OpenAI API endpoint for the batch requests Options: Chat Completions, Embeddings, Completions Default: /v1/chat/completions
Metadata	Object		Custom metadata for tracking the batch job (up to 16 key-value pairs)

Output

Later steps use this data by picking this step in the $ picker. For example, batch_id is $.<step ID>.output.batch_id.

{
  "batch_id": "string",
  "input_file_id": "string",
  "status": "validating"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
batch_id	Text	Unique identifier for the batch job
input_file_id	Text	The ID of the input file used for the batch
status	Text	Current status of the batch job

Retrieve Batch Job

Retrieve the status and details of a batch job by batch ID

Settings

Field	Type	Required	Description
Batch ID	Text	Yes	The ID of the batch job to retrieve

Output

Later steps use this data by picking this step in the $ picker. For example, completion_window is $.<step ID>.output.completion_window.

{
  "completion_window": "24h",
  "created_at": 0,
  "endpoint": "string",
  "error_file_id": "string",
  "errors": {
    "data": [
      {
        "code": "string",
        "line": 0,
        "message": "string",
        "param": "string"
      }
    ],
    "object": "string"
  },
  "expires_at": 0,
  "id": "string",
  "input_file_id": "string",
  "metadata": {},
  "object": "batch",
  "output_file_id": "string",
  "request_counts": {
    "completed": 0,
    "failed": 0,
    "total": 0
  },
  "status": "validating"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
completion_window	Text	The time window for batch completion (e.g., '24h')
created_at	Number	Unix timestamp (in seconds) of when the batch was created
endpoint	Text	The API endpoint that the batch is targeting
error_file_id	Text	The ID of the error file containing processing errors (null if no errors)
errors	Object	Error information if the batch failed
errors.data[]	List	Array of error details
errors.data[].code	Text	Error code
errors.data[].line	Number	Line number where the error occurred
errors.data[].message	Text	Error message
errors.data[].param	Text	Parameter related to the error
errors.object	Text	The object type
expires_at	Number	Unix timestamp (in seconds) when the batch will expire
id	Text	The unique identifier for the batch
input_file_id	Text	The ID of the input file used for the batch
metadata	Object	Custom metadata for tracking the batch job
object	Text	The object type, which is always 'batch'
output_file_id	Text	The ID of the output file containing batch results (null if not completed)
request_counts	Object	Counts of requests by status
request_counts.completed	Number	Number of completed requests
request_counts.failed	Number	Number of failed requests
request_counts.total	Number	Total number of requests in the batch
status	Text	The current status of the batch

Cancel Batch Job

Cancel a batch job that is in progress. The batch will be in status 'cancelling' for up to 10 minutes and then change to 'cancelled' status.

Settings

Field	Type	Required	Description
Batch ID	Text	Yes	The ID of the batch job to cancel

Output

Later steps use this data by picking this step in the $ picker. For example, completion_window is $.<step ID>.output.completion_window.

{
  "completion_window": "24h",
  "created_at": 0,
  "endpoint": "string",
  "error_file_id": "string",
  "errors": {
    "data": [
      {
        "code": "string",
        "line": 0,
        "message": "string",
        "param": "string"
      }
    ],
    "object": "string"
  },
  "expires_at": 0,
  "id": "string",
  "input_file_id": "string",
  "metadata": {},
  "object": "batch",
  "output_file_id": "string",
  "request_counts": {
    "completed": 0,
    "failed": 0,
    "total": 0
  },
  "status": "validating"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
completion_window	Text	The time window for batch completion (e.g., '24h')
created_at	Number	Unix timestamp (in seconds) of when the batch was created
endpoint	Text	The API endpoint that the batch is targeting
error_file_id	Text	The ID of the error file containing processing errors (null if no errors)
errors	Object	Error information if the batch failed
errors.data[]	List	Array of error details
errors.data[].code	Text	Error code
errors.data[].line	Number	Line number where the error occurred
errors.data[].message	Text	Error message
errors.data[].param	Text	Parameter related to the error
errors.object	Text	The object type
expires_at	Number	Unix timestamp (in seconds) when the batch will expire
id	Text	The unique identifier for the batch
input_file_id	Text	The ID of the input file used for the batch
metadata	Object	Custom metadata for tracking the batch job
object	Text	The object type, which is always 'batch'
output_file_id	Text	The ID of the output file containing batch results (null if not completed)
request_counts	Object	Counts of requests by status
request_counts.completed	Number	Number of completed requests
request_counts.failed	Number	Number of failed requests
request_counts.total	Number	Total number of requests in the batch
status	Text	The current status of the batch

Retrieve File

Retrieve metadata and information about a previously uploaded file by its ID

Settings

Field	Type	Required	Description
File ID	Text	Yes	The ID of the file to retrieve

Output

Later steps use this data by picking this step in the $ picker. For example, bytes is $.<step ID>.output.bytes.

{
  "bytes": 0,
  "created_at": 0,
  "expires_at": 0,
  "filename": "string",
  "id": "string",
  "object": "file",
  "purpose": "assistants",
  "status": "uploaded",
  "status_details": "string"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
bytes	Number	Size of the file in bytes
created_at	Number	Unix timestamp (in seconds) of when the file was created
expires_at	Number	Unix timestamp (in seconds) when the file will expire
filename	Text	Name of the uploaded file
id	Text	Unique identifier for the file
object	Text	The object type, which is always 'file'
purpose	Text	Intended purpose of the file
status	Text	Current processing status of the file
status_details	Text	Additional details about the file status (deprecated)

Delete File

Delete a file from OpenAI storage. Returns a deletion status object.

Settings

Field	Type	Required	Description
File ID	Text	Yes	The ID of the file to delete

Output

Later steps use this data by picking this step in the $ picker. For example, deleted is $.<step ID>.output.deleted.

{
  "deleted": false,
  "file_id": "string",
  "object": "string"
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
deleted	Yes / No	Whether OpenAI deleted the file
file_id	Text	ID of the deleted file
object	Text	Object type returned by OpenAI, usually file

Retrieve File Content

Retrieve the actual contents of a file. Returns the file content as a string.

Settings

Field	Type	Required	Description
File ID	Text	Yes	The ID of the file whose content to retrieve

Output

Later steps use this data by picking this step in the $ picker. For example, content is $.<step ID>.output.content.

{
  "content": "string",
  "file_id": "string",
  "size": 0
}

Example values show the shape of the data, not real values. See Using data from earlier steps.

Field descriptions

Field	Type	Description
content	Text	File content as text
file_id	Text	ID of the file
size	Number	Content size in bytes