Upstash QStash SDK

npm (scoped)

Note

> This project is in GA Stage.

The Upstash Professional Support fully covers this project. It receives regular updates, and bug fixes. The Upstash team is committed to maintaining and improving its functionality.

QStash is an HTTP based messaging and scheduling solution for serverless and edge runtimes.

It is 100% built on stateless HTTP requests and designed for:

Serverless functions (AWS Lambda …)
Cloudflare Workers (see the example)
Fastly Compute@Edge
Next.js, including edge
Deno
Client side web/mobile applications
WebAssembly
and other environments where HTTP is preferred over TCP.

How does QStash work?

QStash is the message broker between your serverless apps. You send an HTTP request to QStash, that includes a destination, a payload and optional settings. We durably store your message and will deliver it to the destination API via HTTP. In case the destination is not ready to receive the message, we will retry the message later, to guarentee at-least-once delivery.

Quick Start

Install

npm

npm install @upstash/qstash

Get your authorization token

Go to upstash and copy the token.

Basic Usage:

Publishing a message

import { Client } from "@upstash/qstash";
/**
 * Import a fetch polyfill only if you are using node prior to v18.
 * This is not necessary for nextjs, deno or cloudflare workers.
 */
import "isomorphic-fetch";

const c = new Client({
  token: "<QSTASH_TOKEN>",
});

const res = await c.publishJSON({
  url: "https://my-api...",
  // or urlGroup: "the name or id of a url group"
  body: {
    hello: "world",
  },
});
console.log(res);
// { messageId: "msg_xxxxxxxxxxxxxxxx" }

Receiving a message

How to receive a message depends on your http server. The Receiver.verify method should be called by you as the first step in your handler function.

import { Receiver } from "@upstash/qstash";

const r = new Receiver({
  currentSigningKey: "..",
  nextSigningKey: "..",
});

const isValid = await r.verify({
  /**
   * The signature from the `Upstash-Signature` header.
   *
   * Please note that on some platforms (e.g. Vercel or Netlify) you might
   * receive the header in lower case: `upstash-signature`
   *
   */
  signature: "string";

  /**
   * The raw request body.
   */
  body: "string";
})

Publishing a message to Open AI or any Open AI Compatible LLM

No need for complicated setup your LLM request. We’ll call LLM and schedule it for your serverless needs.

import { Client } from "@upstash/qstash";

const c = new Client({
  token: "<QSTASH_TOKEN>",
});

const result = await client.publishJSON({
  llmProvider: "openai", // We currently support open-ai and together-ai, but QStash will work with any OpenAI compatible API
  llmToken: "YOUR_TOKEN",
  url: "OPEN_AI_COMPATIBLE_BASE_URL",
  body: {
    model: "gpt-3.5-turbo",
    messages: [
      {
        role: "user",
        content: "Where is the capital of Turkey?",
      },
    ],
  },
  callback: "https://oz.requestcatcher.com/",
});

Chatting with your favorite LLM

You can easily start streaming Upstash, OpenAI or TogetherAI responses from your favorite framework(Next.js) or library

const response = await client.chat().create({
  provider: "upstash", // Optionally, provider: "openai"
  model: "meta-llama/Meta-Llama-3-8B-Instruct", // Optionally, model: "gpt-3.5-turbo",
  llmToken: process.env.OPENAI_API_KEY!,
  messages: [
    {
      role: "system",
      content: "from now on, foo is whale",
    },
    {
      role: "user",
      content: "what exactly is foo?",
    },
  ],
  stream: true,
  temperature: 0.5,
});

Docs

See the documentation for details.