Create Chat or Text Completion (OpenAI) v1.0.0 Help ​

Creates a completion for a chat or single message.

How can I use the Step? ​

You can use this Step to automate customer interactions in chatbots, generate dynamic content, and facilitate AI-driven conversations within the Flow. It's suitable for customer service automation, educational tools, and enhancing user engagement on websites or applications.

How does the Step work? ​

The Step utilizes the OpenAI Chat Completions API and operates in two following modes:

  • Chat Completion: Designed for multi-turn conversations and can maintain a conversation history. Uses system and user messages to form an array of message objects to create requests to the Chat Completion API.
  • Text Completion: Tailored for single-turn tasks without conversation, this mode requires only user messages. It bypasses the need for conversation history and system messages, focusing solely on generating a response to the immediate input.

This approach ensures flexibility, allowing the Step to handle both complex, context-driven interactions and straightforward, single-query responses efficiently.

For more details, visit the Chat Completions guide.

Prerequisites ​

  • Before using this Step, ensure you have an OpenAI API key. If you don't have one, obtain it from your API Keys page.

Authorization ​

To authorize the Step, you can inherit authorization from the previous Step or select/create a new authorization in the current step.

To create a new authorization, follow these steps:

  1. For the authorization options, pick Select authorization in current Step.
  2. In the list of authorizations, select the Create a new authorization option.
  3. In the Add authorization modal, do the following:
    • Name your new authorization.
    • Enter your OpenAI API key.
    • Click Add to confirm.

Request settings ​

Configure the request settings as per the Create chat completion endpoint:

  • Mode: Determines whether the Step operates in Chat completion or Text completion mode.
  • System message (only for Chat completion mode): Defines the assistant's behavior, including its personality and any specific instructions for interaction.
  • User message: The text contents of the user message. Contains the requests or comments directed towards the assistant for response.
  • History capacity (only for Chat completion mode): Specifies the number of past messages retained in the conversation history.
  • Model: Identifies the AI model selected for generating responses.
  • Temperature: Controls the level of creativity exhibited in the assistant's responses.
  • Max tokens: The maximum number of tokens in the generated response.
  • Frequency penalty: Encourages the model to generate unique responses.
  • Presence penalty: Encourages the model to stay on-topic.
  • Stop: A sequence where the model should stop generating tokens.
  • Top P: Controls the diversity of the generated response.

Advanced settings ​

Specify additional optional parameters:

  • Stream: Toggle for streaming API response.
  • N: Number of completions to generate.
  • Result example: Defines output structure for accessing specific properties in the Merge field.

Merge field settings ​

The Step returns the result as a JSON object and stores it under the Merge field name. To learn more about Merge fields and how to work with them, see our Merge fields guide.

Output example ​

The output example shows the structure of the generated completion object:

json
{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "gpt-3.5-turbo-0613",
  "system_fingerprint": "fp_44709d6fcb",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "\n\nHello there, how may I assist you today?",
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 12,
    "total_tokens": 21
  }
}
{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "gpt-3.5-turbo-0613",
  "system_fingerprint": "fp_44709d6fcb",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "\n\nHello there, how may I assist you today?",
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 9,
    "completion_tokens": 12,
    "total_tokens": 21
  }
}

For more details, see the Сhat completion object API reference.

Error Handling ​

By default, the Step handles errors using a separate exit. So if any error occurs during the Step execution, the Flow proceeds down the error exit. For more information, see Error and timeout handling.

Reporting ​

The Step automatically generates Reporting events during its execution, allowing for real-time tracking and analysis of its performance and user interactions. To learn more, see Reporting events

Service dependencies ​

  • flow builder - v2.28.3
  • event-manager - v2.3.0
  • deployer - v2.6.0
  • library v2.11.3
  • studio v2.64.1

Release notes ​

v1.0.0

  • Initial release