Search

How to get OpenAI to give you valid JSON.

I was having trouble getting it to work with @std/openai, so you'll need your own OPENAI_API_KEY for this

Code from https://platform.openai.com/docs/guides/structured-outputs/introduction

import OpenAI from "npm:openai";

import { zodResponseFormat } from "npm:openai/helpers/zod";

import { z } from "npm:zod";

const openai = new OpenAI();

const CalendarEvent = z.object({

participants: z.array(z.string()),

const completion = await openai.beta.chat.completions.parse({

model: "gpt-4o-2024-08-06",

Use OpenAI's chat completion API with [`std/openai`](https://www.val.town/v/std/openai). This integration enables access to OpenAI's language models without needing to acquire API keys.

For free Val Town users, [all calls are sent to `gpt-4o-mini`](https://www.val.town/v/std/openaiproxy?v=12#L85).

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

get chat(): RawOpenAI["beta"]["chat"] {

return this.rawOpenAIClient.beta.chat;

Use OpenAI's chat completion API with [`std/openai`](https://www.val.town/v/std/openai). This integration enables access to OpenAI's language models without needing to acquire API keys.

For free Val Town users, [all calls are sent to `gpt-4o-mini`](https://www.val.town/v/std/openaiproxy?v=12#L85).

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

get chat(): RawOpenAI["beta"]["chat"] {

return this.rawOpenAIClient.beta.chat;

import { OpenAI } from "https://esm.town/v/std/openai";

const openai = new OpenAI();

export default async function(req: Request): Promise<Response> {

const messages = prepareMessages(selection);

log.info("Prepared messages for OpenAI");

const openai = new OpenAI();

const stream = await openai.chat.completions.create({

messages,

stream: true,

// log.info("OpenAI stream created");

return streamResponse(stream);

if (content) {

// log.debug("Received chunk from OpenAI:", content);

const encodedChunk = encoder.encode(JSON.stringify({ chunk: content }) + "\n");

import { OpenAI } from "https://esm.town/v/std/openai";

export default async function(req: Request): Promise<Response> {

const openai = new OpenAI();

const stream = await openai.chat.completions.create({

stream: true,

Use OpenAI's chat completion API with [`std/openai`](https://www.val.town/v/std/openai). This integration enables access to OpenAI's language models without needing to acquire API keys.

For free Val Town users, [all calls are sent to `gpt-4o-mini`](https://www.val.town/v/std/openaiproxy?v=12#L85).

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

get chat(): RawOpenAI["beta"]["chat"] {

return this.rawOpenAIClient.beta.chat;

This val uses your OpenAI token if you have one, and the @std/openai if not, so it provides limited OpenAI usage for free.

import { chat } from "https://esm.town/v/stevekrouse/openai";

import type { ChatCompletion, ChatCompletionCreateParamsNonStreaming, Message } from "npm:@types/openai";

async function getOpenAI() {

if (Deno.env.get("OPENAI_API_KEY") === undefined) {

const { OpenAI } = await import("https://esm.town/v/std/openai");

return new OpenAI();

const { OpenAI } = await import("npm:openai");

return new OpenAI();

* Initiates a chat conversation with OpenAI's GPT model and retrieves the content of the first response.

const openai = await getOpenAI();

const completion = await openai.chat.completions.create(createParams);

Get started using OpenAI's chat completion without the need to set your own API keys.

Here's a quick example to get you started with the Val Town OpenAI wrapper:

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

return this.rawOpenAIClient.beta;

import { OpenAI } from "npm:openai";

const openai = new OpenAI();

const functionExpression = await openai.chat.completions.create({

"messages": [

import process from "node:process";

export const openaiCompletion = async (prompt) => {

const { OpenAI } = await import("https://deno.land/x/openai/mod.ts");

const openAI = new OpenAI(process.env.OPENAI_API_KEY);

const completion = openAI.createCompletion({

model: "text-davinci-003",

Use OpenAI's chat completion API with [`std/openai`](https://www.val.town/v/std/openai). This integration enables access to OpenAI's language models without needing to acquire API keys.

import { OpenAI } from "https://esm.town/v/std/openai";

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

get chat(): RawOpenAI["beta"]["chat"] {

return this.rawOpenAIClient.beta.chat;

Use OpenAI's chat completion API with [`std/openai`](https://www.val.town/v/std/openai). This integration enables access to OpenAI's language models without needing to acquire API keys.

For free Val Town users, [all calls are sent to `gpt-3.5-turbo`](https://www.val.town/v/std/openaiproxy?v=5#L69).

import { type ClientOptions, OpenAI as RawOpenAI } from "npm:openai";

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

export class OpenAI {

private rawOpenAIClient: RawOpenAI;

* API Client for interfacing with the OpenAI API. Uses Val Town credentials.

this.rawOpenAIClient = new RawOpenAI({

baseURL: "https://std-openaiproxy.web.val.run/v1",

return this.rawOpenAIClient.chat;

get chat(): RawOpenAI["beta"]["chat"] {

return this.rawOpenAIClient.beta.chat;

import { cors } from 'npm:hono/cors';

import { OpenAI } from "npm:openai";

const app = new Hono();

const openai = new OpenAI();

app.use('*', cors({

<head>

<title>OpenAI Prompt Example</title>

<script defer src="https://cdn.jsdelivr.net/npm/alpinejs@3.x.x/dist/cdn.min.js"></script>

<div class="container mx-auto py-8">

<h1 class="text-4xl font-bold mb-4">OpenAI Prompt Example</h1>

<form action="/prompt" method="GET">

try {

const response = await openai.chat.completions.create({

model: "gpt-4",

} catch (error) {

console.error('OpenAI API error:', error);

return c.redirect('/?response=Error%20occurred.');

This val is a proxy server that interacts with the OpenAI API to generate responses based on prompts in the request body. The function handles incoming HTTP POST requests, processes the prompt, and returns a response generated by the LLM.

- Server-side: (Optional) An active OpenAI API key

- Client-side: Something that can make POST requests (browser code, Postman, cURL, another Val, etc)

curl -X POST https://ashryanio-openaiproxy.web.val.run -H "Content-Type: application/json" -d '{"prompt": "Hello, OpenAI!"}'

- **Content-Type**: application/json

- **Body**: JSON object containing the response from the OpenAI language model.

import { OpenAI } from "https://esm.town/v/std/openai";

const openai = new OpenAI();

* Handles the incoming request and returns a response.

statusText: "Method Not Allowed",

* Gets the response from the OpenAI language model.

* @param {string} prompt - The prompt for the language model.

async function getLlmResponse(prompt: string) {

const completion = await openai.chat.completions.create({

"messages": [

import { fetchJSON } from "https://esm.town/v/stevekrouse/fetchJSON";

export function openaiFineTune({ key, model, trainingFile }: {

key: string;

model?: string;

trainingFile: string;

return fetchJSON(

"https://api.openai.com/v1/fine_tuning/jobs",

method: "POST",

body: JSON.stringify({