Interface OpenAiClient

A client for interacting with the OpenAI API.

See

https://beta.openai.com/docs/quickstart

Remarks

This client provides methods for calling the OpenAI API's endpoints for text completion, language generation, and other tasks.

To create an OpenAiClient, use the HttpOpenAiClient.create method.

The OpenAiClient's methods return Promises that resolve to the JSON response returned by the OpenAI API. If the OpenAI API returns an error, the Promise will be rejected with an instance of an HttpError that includes the error message and HTTP status code.

openai.getCompletion(...)
  .then((response) => {
    console.log(response);
  })
  .catch((error) => {
    console.error(error);
  });

Hierarchy

  • OpenAiClient

Implemented by

Methods

getCompletion getEdit getUrl

Methods

getCompletion

  • Calls the OpenAI APIs text completion endpoint to generate text based on the given prompt.

    Default values for the optional parameters are selected based on the model.

    Returns

    • A promise that resolves to the response from the OpenAI API.

    Throws

    • If the OpenAI API returns an error.

    Throws

    • If the OpenAI API returns a response in an unexpected format.

    Parameters

    • prompt: string

      The prompt to use for text completion.

    • Optional model: string

      The OpenAI API model to use for text completion.

    • Optional max_tokens: number

      The maximum number of tokens (words and punctuation) to generate in the completion.

    • Optional temperature: number

      Controls the "creativity" of the completion. A higher value means the model will take more risks.

    • Optional top_p: number

      Controls the "confidence" of the completion. A lower value means the model will be more confident in its words.

    • Optional frequency_penalty: number

      Controls the "diversity" of the completion. A higher value means the model will avoid repetition of words.

    • Optional presence_penalty: number

      Controls the "relevance" of the completion. A higher value means the model will try to match the prompt more closely.

    Returns Promise<OpenAiCompletionResponseDTO>

getEdit

  • Calls the OpenAI APIs text edit endpoint to generate text based on the given input and instruction.

    Default values for the optional parameters are selected based on the model.

    Returns

    • A promise that resolves to the response from the OpenAI API.

    Throws

    • If the OpenAI API returns an error.

    Throws

    • If the OpenAI API returns a response in an unexpected format.

    Parameters

    • instruction: string

      The instruction to use for text editing.

    • Optional input: string

      The input to use for text editing.

    • Optional model: string

      The OpenAI API model to use for text completion.

    • Optional n: number
    • Optional temperature: number

      Controls the "creativity" of the completion. A higher value means the model will take more risks.

    • Optional top_p: number

      Controls the "confidence" of the completion. A lower value means the model will be more confident in its words.

    Returns Promise<OpenAiEditResponseDTO>

getUrl

Settings

Member Visibility

Theme

Generated using TypeDoc