This library exports OpenAI API calls through the OpenAI class. Refer to the OpenAI API documentation for more details.
Instantiate the OpenAI class with your API key, then call the desired method:
import { OpenAI } from "https://deno.land/x/openai_api/mod.ts";
const openai = new OpenAI("your-api-key-here");
const textCompletion = await openai.text.complete({
model: "davinci",
prompt: "Once upon a time",
maxTokens: 50,
n: 1,
});
console.log(textCompletion.choices[0].text);
The above generates a short text completion based on the prompt "Once upon a time" using the Davinci engine.
The following APIs are currently available:
Audio:
- openai.audio.transcription(request): Transcribe audio to input language. (Audio to Text)
- openai.audio.translation(request): Translates audio into into English.
Chat:
- openai.chat.complete(request): Return a chat completion response.
Embeddings:
- openai.embedding.create(request): Create embedding vector representing the input text.
Files:
- openai.files.list(): Return files that belong to user's organization.
- openai.files.upload(file): Upload file to be used for OpenAI features.
- openai.files.delete(fileId): Delete file.
- openai.files.info(fileId): Retrieve information about a file.
- openai.files.contents(fileId): Retrieve contents of a file.
Fine-Tuning:
- openai.fineTune.list(): Lists fine-tuned models.
- openai.fineTune.create(fileId, model): Creates a job to fine-tunes a specified model.
- openai.fineTune.delete(fineTuneId): Deletes a fine-tune model.
- openai.fineTune.info(fineTuneId): Gets info about the fine-tune job.
- openai.fineTune.cancel(fineTuneId): Cancels a fine-tune job.
- openai.fineTune.listEvents(fineTuneId): Get fine-grained status updates for a fine-tune job.
Image:
- openai.image.create(request): Generate a new image.
- openai.image.edit(request): Creates an edited or extended image with an original image and prompt.
- openai.image.variation(request): Creates a variation of a given image.
Models:
- openai.models.list(): Lists currently available models, provides basic info.
- openai.models.retrieve(modelId): Retrieves a model instance, provides basic info.
Moderation:
- openai.moderation.check(request): Classifies if text violates OpenAI's Content Policy
Text:
- openai.text.complete(request): Model returns one or more predicted completions based on prompt.
- openai.text.edit(request): Creates a new edit for the provided input, instruction, and parameters.
This library is licensed under the MIT License.