Skip to main content

OpenAI Component

Interact with OpenAI's models and build AI Agents

Component key: openai

Changelog ↓

Description​

OpenAI is an artificial intelligence research laboratory that develops AI models and APIs. It has produced models including GPT-4 and GPT-3.5 Turbo for text generation, and DALL·E 3 for image generation.

This component wraps some of OpenAI's REST API endpoints. Additional endpoints can be referenced in their API documentation. You can use the Raw Request action to interact with endpoints that are not covered by other actions.

AI Agents​

The agent system is built on the OpenAI Agents SDK and provides a visual way to create conversational AI agents with tools and multi agent workflows. Build AI agents using OpenAI's Agents SDK with Prismatic's drag and drop Designer interface. Create agents with custom instructions, tools, and multi agent workflows without writing code.

Quick Start​

Create a simple AI assistant in three steps:

  1. Add "Agent: Create" action from the OpenAI component

    • Set Name: "Assistant"
    • Set Instructions: "You are a helpful assistant"
    • Select Model: "gpt-4o-mini" (fast and cost-effective)

  2. Add "Agent: Run" action

    • Set OpenAI Connection: Select your configured connection
    • Set Agent Configuration: Reference the Create Agent output: {{$agentCreate.results}}
    • Set User Input: "What is the weather in San Francisco?"

  3. Test your flow - The agent's response will be in {{$agentRun.results.finalOutput}}

Core Concepts​

Agent Configuration​

Agents are configured and executed in two steps:

  • Agent Configuration: Created using the "Agent: Create" action, which outputs a configuration object defining the agent's behavior, tools, and capabilities
  • Agent Execution: The "Agent: Run" action executes the configured agent with user input and returns the response

Tools​

Tools extend agent capabilities:

  • Flow Tools: Let agents trigger other Prismatic flows
  • Approval Tools: Require human approval before proceeding
  • Built in Tools: Web search, code interpreter, file search, and image generation

MCP Servers​

Model Context Protocol servers connect agents to external systems:

  • Local MCP: Run tools over STDIO
  • Remote MCP: Connect to remote HTTPS servers

Configuration Examples​

Understanding the configuration objects helps you work with agent actions.

Tool Configurations​

Flow Tool - Output from "Agent: Create Flow Tool" action:

{
  "type": "flow",
  "tool": {
    "type": "function",
    "function": {
      "name": "send_email",
      "description": "Send a welcome email to a new customer",
      "parameters": {
        "type": "object",
        "properties": {
          "to": { "type": "string" },
          "subject": { "type": "string" },
          "body": { "type": "string" }
        }
      }
    }
  },
  "toolName": "send_email"
}

Approval Tool - Output from "Agent: Create Human Approval Tool" action:

{
  "type": "approval",
  "tool": {
    "type": "function",
    "function": {
      "name": "delete_confirmation",
      "description": "Confirm before deleting records",
      "needsApproval": true
    }
  },
  "toolName": "delete_confirmation"
}

MCP Server Configurations​

Local MCP Server - Output from "Agent: Add Local MCP Server" action:

{
  "mcpServer": {
    "type": "stdio",
    "label": "sequential-thinking",
    "command": "npx",
    "args": ["-y", "@modelcontextprotocol/server-sequential-thinking"]
  }
}

Remote MCP Server - Output from "Agent: Add Remote MCP Server" action:

{
  "mcpServer": {
    "type": "http",
    "label": "company-data-server",
    "url": "https://api.company.com/mcp",
    "headers": {
      "Authorization": "Bearer YOUR_TOKEN"
    }
  }
}

Agent Configurations​

Simple Agent:

{
  "agent": {
    "name": "Assistant",
    "instructions": "You are a helpful assistant",
    "model": "gpt-4o-mini"
  }
}

Agent with Structured Output:

{
  "agent": {
    "name": "Email Classifier",
    "instructions": "Classify emails by priority and sentiment",
    "model": "gpt-4o",
    "outputType": {
      "type": "json_schema",
      "name": "EmailClassification",
      "strict": true,
      "schema": {
        "type": "object",
        "properties": {
          "priority": {
            "type": "string",
            "enum": ["high", "medium", "low"]
          },
          "sentiment": {
            "type": "string",
            "enum": ["positive", "negative", "neutral"]
          }
        }
      }
    }
  }
}

Common Workflows​

Simple Q&A Bot​

  1. Add Agent: Create action

    • Name: "Q&A Bot"
    • Instructions: "Answer questions about our product. Be helpful and concise."
    • Model: "gpt-4o-mini"

  2. Add Agent: Run action

    • Agent Configuration: {{$agentCreate.results}}
    • User Input: {{$trigger.body.question}}

  3. Add Webhook Reply action

    • Response Body: {{$agentRun.results.finalOutput}}

Message Triage with Structured Output​

  1. Add Agent: Create action with output schema:

    {
      "type": "object",
      "properties": {
        "priority": {
          "type": "string",
          "enum": ["high", "medium", "low"]
        },
        "sentiment": {
          "type": "string",
          "enum": ["positive", "negative", "neutral"]
        },
        "summary": {
          "type": "string"
        }
      },
      "required": ["priority", "sentiment", "summary"]
    }

  2. Add Agent: Run action

    • User Input: {{$trigger.body.emailContent}}

  3. Use the structured output:

    • Priority: {{$agentRun.results.finalOutput.priority}}
    • Sentiment: {{$agentRun.results.finalOutput.sentiment}}

Multi Agent Customer Support​

  1. Create specialist agents:

    • Billing Agent: "Handle billing and payment questions"
    • Technical Agent: "Solve technical problems"

  2. Create main triage agent:

    • Instructions: "Route questions to the right specialist"

  3. Run with handoffs:

    • Agent Configuration: {{$triageAgent.results}}
    • Handoffs: [{{$billingAgent.results}}, {{$techAgent.results}}]

Human Approval Workflow​

  1. Add Agent: Create Human Approval Tool

    • Name: "Delete Confirmation"

  2. Add Agent: Create with tools:

    • Tools: {{$approvalTool.results}}

  3. Add Agent: Run action

  4. Add Branch on Expression:

    • If {{$agentRun.results.hasInterruptions}} is true:
      • Send approval request
      • Add Agent: Resume Run with approval response

Multi Turn Conversation​

  1. Initial message:

    • Add Agent: Run with user input
    • Store: {{$firstRun.results.history}}

  2. Follow up message:

    • Add Agent: Run with same configuration
    • History: {{$firstRun.results.history}}
    • Agent will maintain conversation context

Flow Tool Integration​

  1. Add Agent: Create Flow Tool

    • Select existing flow (e.g., "Send Email")
    • Tool Description: "Send emails to customers"

  2. Add Agent: Create with tools:

    • Tools: {{$flowTool.results}}

  3. Agent will invoke your flow when appropriate

Using Configurations in Designer​

When actions output configurations, reference them in subsequent actions:

  • Tool outputs: {{$createFlowTool.results}} or {{$approvalTool.results}}
  • MCP Server outputs: {{$addLocalMcp.results}} or {{$addRemoteMcp.results}}
  • Agent configuration: {{$createAgent.results}}

Pass multiple tools/servers as a list:

  • Tools: [{{$tool1.results}}, {{$tool2.results}}]
  • MCP Servers: [{{$mcp1.results}}, {{$mcp2.results}}]

File Processing with Agents​

The OpenAI component supports file uploads and processing, allowing agents to work with documents, images, and other files.

Uploading Files​

Use the File: Upload action to upload files to OpenAI:

  1. Add File: Upload action

    • File: Your file data (binary or base64)
    • Filename: "document.pdf"
    • Purpose: "assistants" (for use with agents)

  2. Store the file ID: {{$uploadFile.results.id}}

Processing Files with Agents​

Pass file IDs to agents for processing PDFs, images, and other documents:

  1. Upload your files first:

    File: Upload → Returns file ID

  2. Add Agent: Create action:

    • Instructions: "Analyze the provided documents and extract key information"
    • Model: "gpt-4o" (for best document understanding)

  3. Add Agent: Run action:

    • Agent Configuration: {{$agentCreate.results}}
    • User Input: "Summarize the main points from the attached document"
    • File IDs: [{{$uploadFile.results.id}}]

Common File Processing Workflows​

PDF Analysis:

  1. Upload PDF with File: Upload
  2. Create agent with document analysis instructions
  3. Run agent with file ID to extract text, tables, or answer questions

Image Processing:

  1. Upload image with File: Upload
  2. Create agent with vision capabilities (gpt-4o model)
  3. Run agent with file ID to describe, analyze, or extract text from images

Multi-File Processing:

  1. Upload multiple files
  2. Pass array of file IDs: [{{$file1.results.id}}, {{$file2.results.id}}]
  3. Agent can reference and compare across multiple documents

Connections​

OpenAI API Key​

Integrations can authenticate with OpenAI using API keys. Generate an API key at platform.openai.com/account/api-keys.

If your user is associate with one organization, you can leave the connection's organization field blank. Otherwise, specify your organization's ID.

InputNotesExample
API Key

Generate an API key at https://platform.openai.com/account/api-keys

Organization

Your organization ID. If you have just one organization, this value is optional.

org-9AeeV8Yj8WK9tW6QEZNHDI8Z

Triggers​

Tool Flow Trigger​

Marks this flow as a tool that can be called by AI agents | key: toolFlowTrigger

Data Sources​

List Models​

Lists the currently available models, and provides basic information about each one such as the owner and availability. | key: listModels | type: picklist

InputNotesExample
ConnectionOpenAI Connection
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "result": [
    {
      "label": "gpt-4-turbo-preview",
      "key": "gpt-4-turbo-preview"
    },
    {
      "label": "gpt-4",
      "key": "gpt-4"
    },
    {
      "label": "gpt-3.5-turbo",
      "key": "gpt-3.5-turbo"
    },
    {
      "label": "dall-e-3",
      "key": "dall-e-3"
    }
  ]
}

Actions​

Agent: Add Local MCP Server​

Configure a local MCP server that runs via command line | key: addLocalMcpServer

InputNotesExample
Arguments

Arguments to pass to the command

-y
Command

Command to execute the MCP server (currently only npx is supported)

npx
Server Label

Unique identifier for the MCP server

sequential-thinking
Tool Filter

List of specific tool names to expose. If empty, all tools are exposed.

think_forward

Agent: Add Remote MCP Server​

Configure a remote MCP server accessed via HTTP | key: addRemoteMcpServer

InputNotesExample
Headers

HTTP headers for authentication or other needs. Key = header name, Value = header value

Authorization -> Bearer YOUR_TOKEN
Server Label

Unique identifier for the MCP server

company-data-server
Server URL

URL endpoint for the MCP server

https://api.company.com/mcp
Tool Filter

List of specific tool names to expose. If empty, all tools are exposed.

query_database

Agent: Classify and Branch​

Use AI to analyze input and route to the appropriate branch | key: classifyAndBranch

InputNotesExample
MCP Servers

List of MCP server configurations. Each should be the output from an Add MCP Server action.

Tools

List of tools available to the agent. Each tool should be the output from a Create Tool action.

Branch Definitions

Define branches and their classification criteria. Key = branch name, Value = description of when to use this branch

needs_reply -> Email requires a response or action informational -> FYI only, no action needed spam -> Unsolicited commercial email urgent -> Time-sensitive, requires immediate attention
Classification Instructions

Additional instructions for how to classify the input

Consider the sender, subject urgency, and whether a response is explicitly requested. Prioritize 'urgent' for anything mentioning deadlines today or tomorrow.
File IDs

File IDs from previously uploaded files to OpenAI. Use the Upload File action to get these IDs.

file-abc123
Input Text

The text to analyze and classify (e.g., email content)

Can you make the urgent meeting tomorrow morning at 10am?
Model

Select an OpenAI model to use for the request. The list of available models will be fetched from your OpenAI account.

gpt-5-mini-2025-08-07
OpenAI Connection

Connection to OpenAI API with your API key

Agent: Create​

Create an AI agent with customizable instructions. | key: createAgent

InputNotesExample
Handoff Description

Description for handoff routing. Used by other agents to understand when to hand off to this agent.

Handles technical support questions
Instructions

System instructions that define the agent's behavior and capabilities

You are a helpful customer support agent. Answer questions politely and accurately.
MCP Servers

List of MCP server configurations. Each should be the output from an Add MCP Server action.

Model

Select an OpenAI model to use for the request. The list of available models will be fetched from your OpenAI account.

gpt-5-mini-2025-08-07
Name

The name of the agent

Customer Support Agent
Output Schema

JSON Schema for structured output. Forces the agent to respond with data matching this schema.

Output Schema Name

Name for the output schema

response
Output Schema Strict

If true, enforces strict schema validation

true
Tools

List of tools available to the agent. Each tool should be the output from a Create Tool action.

Agent: Create Code Interpreter Tool​

Create a code interpreter tool that allows agents to execute Python code | key: createCodeInterpreterTool

InputNotesExample
Container ID

Specific container ID (only used if containerType is 'specific')

Container Type

Container type for code execution

auto
File IDs

Comma-separated list of file IDs to include in the container

Tool Name

Custom name for the code interpreter tool

Code Interpreter

Agent: Create File Search Tool​

Create a file search tool that searches through vector stores | key: createFileSearchTool

InputNotesExample
Include Search Results

Include search results in the LLM output

true
Max Results

Maximum number of results to return

Tool Name

Custom name for the file search tool

File Search
Ranking Algorithm

Algorithm for ranking search results

Score Threshold

Minimum score threshold for results (0-1)

Vector Store IDs

Comma-separated list of vector store IDs to search

Agent: Create Flow Tool​

Create a tool configuration that allows an agent to invoke another flow | key: createFlowTool

InputNotesExample
Flow Name

The flow that this tool will invoke when called by an agent

Requires Approval

If true, the tool will require human approval before the agent can execute the flow

false
Strict Mode

If true, requires exact parameter matching. If false, allows flexibility in agent inputs.

false
Tool Description

Description that helps the agent understand when to use this tool

Send a welcome email to a new customer with their account details

Agent: Create Human Approval Tool​

Create a tool that requires human approval before the agent can proceed | key: createHumanApprovalTool

InputNotesExample
Tool Description

Description that helps the agent understand when to use this tool

Request human approval before proceeding with this action
Tool Name

Name for the approval tool

Human Approval
Parameters Schema

Optional JSON schema for parameters the agent should provide when calling this tool. If not provided, the tool will accept a 'reason' string parameter.

Agent: Create Image Generation Tool​

Create an image generation tool that allows agents to generate images | key: createImageGenerationTool

InputNotesExample
Background

Background type for generated images

auto
Input Fidelity

Input fidelity level

Model

Model to use for generation

gpt-image-1
Moderation

Moderation level for content filtering

auto
Tool Name

Custom name for the image generation tool

Image Generation
Output Compression

Compression level 0-100 (higher = more compression)

Output Format

Image output format

png
Partial Images

Number of partial images to generate

Quality

Quality level for generated images

auto
Size

Image dimensions

auto

Agent: Create Web Search Tool​

Create a web search tool that allows agents to search the web | key: createWebSearchTool

InputNotesExample
User City

City for location-aware searches

User Country

Country for location-aware searches

Tool Name

Custom name for the web search tool

Web Search
User Region

Region/state for location-aware searches

Search Context Size

Amount of context to return from searches

medium
User Timezone

Timezone for time-sensitive searches (e.g., 'America/New_York')

Agent: Resume Run​

Resume an interrupted agent run with approval decisions | key: resumeRun

InputNotesExample
Agent Configuration

Agent configuration object from Create Agent action. Should be the same configuration used in the original run.

Approval Responses

Array of approval responses with updated approvalRequest objects containing functionId and approved status.

Handoff Agents

Same handoff agents configuration used in the original run, if any.

Max Turns

Maximum number of conversation turns the agent can take from this point. Prevents infinite loops.

20
OpenAI Connection

Connection to OpenAI API with your API key

Run State

State from the interrupted run, containing the execution context and interruptions.

Agent: Run​

Run an agent | key: runAgent

InputNotesExample
Agent Configuration

Agent configuration object from Create Agent action

File IDs

File IDs from previously uploaded files to OpenAI. Use the Upload File action to get these IDs.

file-abc123
Handoff Agents

List of agent configurations that can be handed off to. Each should be the output from a Create Agent action.

History

Previous conversation history. Use this to continue a conversation from a previous run. Should be the history array from a previous agent run result.

Max Turns

Maximum number of conversation turns the agent can take. Prevents infinite loops.

20
OpenAI Connection

Connection to OpenAI API with your API key

Previous Response ID

ID of the previous response to continue the conversation. Alternative to passing full history array.

resp_abc123
User Input

The message or question to send to the agent

What is the weather like today?

Create Chat Completion​

Create a completion for the chat message | key: createChatCompletion

InputNotesExample
ConnectionOpenAI Connection
Messages
Model

Select an OpenAI model to use for the request. The list of available models will be fetched from your OpenAI account.

gpt-5-mini-2025-08-07
Number of choices

How many chat completion choices to generate for each input message.

1
Temperature

What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.

0.7
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000
top_p

An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered.

0.9

{
  "data": {
    "id": "chatcmpl-6vf993Onyn1Bwq67AYDTR6fIjzvBu",
    "object": "chat.completion",
    "created": 1679200883,
    "model": "gpt-3.5-turbo-0301",
    "usage": {
      "prompt_tokens": 56,
      "completion_tokens": 21,
      "total_tokens": 77
    },
    "choices": [
      {
        "message": {
          "role": "assistant",
          "content": "The 2020 World Series was played at the Globe Life Field in Arlington, Texas, USA."
        },
        "finish_reason": "stop",
        "index": 0
      }
    ]
  }
}

Create Image​

Create image(s) given a prompt | key: createImage

InputNotesExample
ConnectionOpenAI Connection
Image Size

The size of the generated images. Must be one of 256x256, 512x512, or 1024x1024.

1024x1024
Number of Images

The number of images to generate. Must be between 1 and 10.

1
Prompt

A text description of the desired image(s). The maximum length is 1000 characters.

A cute baby sea otter
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "created": 1589478378,
    "data": [
      {
        "url": "https://..."
      },
      {
        "url": "https://..."
      }
    ]
  }
}

Create Response​

Create a response using the responses endpoint | key: createFunctionCallingResponse

InputNotesExample
ConnectionOpenAI Connection
Input

The input text to process

Model

Select an OpenAI model to use for the request. The list of available models will be fetched from your OpenAI account.

gpt-5-mini-2025-08-07
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "id": "resp-123",
    "object": "response",
    "created": 1677652288,
    "model": "gpt-4.1",
    "choices": [
      {
        "index": 0,
        "message": {
          "role": "assistant",
          "content": "Here is the response to your input."
        },
        "finish_reason": "stop"
      }
    ],
    "usage": {
      "prompt_tokens": 82,
      "completion_tokens": 18,
      "total_tokens": 100
    }
  }
}

Delete File​

Delete a previously uploaded file | key: deleteFile

InputNotesExample
ConnectionOpenAI Connection
File ID

The ID of the file to delete

file-abc123
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "id": "file-abc123",
    "object": "file",
    "deleted": true
  }
}

Get Model by ID​

Get model by ID | key: getModelById

InputNotesExample
ConnectionOpenAI Connection
Model

Select an OpenAI model to use for the request. The list of available models will be fetched from your OpenAI account.

gpt-5-mini-2025-08-07
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "id": "davinci",
    "object": "model",
    "created": 1649359874,
    "owned_by": "openai",
    "permission": [
      {
        "id": "modelperm-U6ZwlyAd0LyMk4rcMdz33Yc3",
        "object": "model_permission",
        "created": 1669066355,
        "allow_create_engine": false,
        "allow_sampling": true,
        "allow_logprobs": true,
        "allow_search_indices": false,
        "allow_view": true,
        "allow_fine_tuning": false,
        "organization": "*",
        "group": null,
        "is_blocking": false
      }
    ],
    "root": "davinci",
    "parent": null
  }
}

List Files​

List previously uploaded files | key: listFiles

InputNotesExample
ConnectionOpenAI Connection
Purpose

Filter files by purpose

Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "object": "list",
    "data": [
      {
        "id": "file-abc123",
        "object": "file",
        "bytes": 120000,
        "created_at": 1677610602,
        "filename": "training_data.jsonl",
        "purpose": "fine-tune",
        "status": "processed",
        "status_details": null
      },
      {
        "id": "file-xyz456",
        "object": "file",
        "bytes": 250000,
        "created_at": 1677610702,
        "filename": "validation_data.jsonl",
        "purpose": "fine-tune",
        "status": "processed",
        "status_details": null
      }
    ]
  }
}

List Models​

List all available models | key: listModels

InputNotesExample
ConnectionOpenAI Connection
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "object": "list",
    "data": [
      {
        "id": "davinci",
        "object": "model",
        "created": 1649359874,
        "owned_by": "openai",
        "permission": [
          {
            "id": "modelperm-U6ZwlyAd0LyMk4rcMdz33Yc3",
            "object": "model_permission",
            "created": 1669066355,
            "allow_create_engine": false,
            "allow_sampling": true,
            "allow_logprobs": true,
            "allow_search_indices": false,
            "allow_view": true,
            "allow_fine_tuning": false,
            "organization": "*",
            "group": null,
            "is_blocking": false
          }
        ],
        "root": "davinci",
        "parent": null
      }
    ]
  }
}

Raw Request​

Send raw HTTP request to OpenAI | key: rawRequest

InputNotesExample
ConnectionOpenAI Connection
Data

The HTTP body payload to send to the URL.

{"exampleKey": "Example Data"}
Debug Request

Enabling this flag will log out the current request.

false
File Data

File Data to be sent as a multipart form upload.

[{key: "example.txt", value: "My File Contents"}]
File Data File Names

File names to apply to the file data inputs. Keys must match the file data keys above.

Form Data

The Form Data to be sent as a multipart form upload.

[{"key": "Example Key", "value": new Buffer("Hello World")}]
Header

A list of headers to send with the request.

User-Agent: curl/7.64.1
Max Retry Count

The maximum number of retries to attempt. Specify 0 for no retries.

0
Method

The HTTP method to use.

Query Parameter

A list of query parameters to send with the request. This is the portion at the end of the URL similar to ?key1=value1&key2=value2.

Response Type

The type of data you expect in the response. You can request json, text, or binary data.

json
Retry On All Errors

If true, retries on all erroneous responses regardless of type. This is helpful when retrying after HTTP 429 or other 3xx or 4xx errors. Otherwise, only retries on HTTP 5xx and network errors.

false
Retry Delay (ms)

The delay in milliseconds between retries. This is used when 'Use Exponential Backoff' is disabled.

0
Timeout

The maximum time that a client will await a response to its request

2000
URL

Input the path only (/v1/images/generations), The base URL is already included (https://api.openai.com). For example, to connect to https://api.openai.com/v1/images/generations, only /v1/images/generations is entered in this field.

/v1/images/generations
Use Exponential Backoff

Specifies whether to use a pre-defined exponential backoff strategy for retries. When enabled, 'Retry Delay (ms)' is ignored.

false

Retrieve File​

Retrieve information about a specific file | key: retrieveFile

InputNotesExample
ConnectionOpenAI Connection
File ID

The ID of the file to retrieve

file-abc123
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "id": "file-abc123",
    "object": "file",
    "bytes": 120000,
    "created_at": 1677610602,
    "filename": "training_data.jsonl",
    "purpose": "fine-tune",
    "status": "processed",
    "status_details": null
  }
}

Upload File​

Upload a file to OpenAI that can be used with various features | key: uploadFile

InputNotesExample
ConnectionOpenAI Connection
File

The file to upload (binary data or base64 encoded)

Filename

Name of the file including extension

training_data.jsonl
Purpose

The intended purpose of the uploaded file

assistants
Timeout (ms)

The maximum amount of time (in MS) to wait for a response.

10000

{
  "data": {
    "id": "file-abc123",
    "object": "file",
    "bytes": 120000,
    "created_at": 1677610602,
    "filename": "training_data.jsonl",
    "purpose": "fine-tune",
    "status": "processed",
    "status_details": null
  }
}

Changelog​

2025-08-23​

Added file operations and enhanced agent capabilities:

  • New file management actions: Upload, Download, and Delete files for use with OpenAI
  • Tool Flow Trigger to enable flows as callable AI agent tools
  • File attachment support in agent runs for processing PDFs, images, and documents
  • Enhanced agent branching with file IDs and MCP server integration
  • Improved agent configuration with custom tools and MCP servers
  • Comprehensive test coverage for new file operations

2025-08-19​

Added AI Agents functionality with support for building agents, custom and OpenAI hosted tools, and MCP server connections.

2025-04-25​

Modernized OpenAI component with updated libraries and upstream API changes.