Node.js Replicate Example

This repo contains a "hello world" example demonstrating the how to prompt a machine learning model hosted by Replicate via node.js. This example is for the Programming from A to Z class at ITP, NYU. (Students in the class, contact me for your API key!)

import Replicate from 'replicate';
import * as dotenv from 'dotenv';
dotenv.config();

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

const version = '8e6975e5ed6174911a6ff3d60540dfd4844201974602551e10e9e87ab143d81e';
const model = 'meta/llama-2-7b-chat';
go();

async function go() {
  const input = {
    prompt: 'Help me with a thesis idea.',
    system_prompt: 'You are a student at ITP, Tisch School of the Arts, NYU.',
  };
  const output = await replicate.run(`${model}:${version}`, { input });
  console.log(output.join('').trim());
}

Replicate

Replicate is a platform for running machine learning models in the cloud from your own code.

Here are a few links to help you explore what is possible with generative AI models.

Getting Started

• Getting started with Replicate in Node.js
• Explore all models

NOTE: models prefixed with ✨ are currently popular in the community.

Language Models

• language
  • llama-2-7b-chat (small)
  • llama-2-13b-chat (medium)
  • ✨ llama-2-70b-chat (large)
• code
  • codellama-7b-instruct (small)
  • codellama-13b-instruct (medium)
  • ✨ codellama-34b-instruct (large)

Images

• ✨ SDXL
• ✨ SDXL fine-tunes
  • make it emoji! sdxl-emoji
  • make it tron! sdxl-tron
  • make it 70s wool! sdxl-woolitize
• Stable Diffusion 2.1
• ✨ illusion
• generate tileable images with material_stable_diffusion

Audio

• For music, try MusicGen or riffusion
• For voice, try bark or xtts
• For voice transcription, try whisperx

Video

• ✨ Animate Diff
• ✨ Hotshot XL
• Zeroscope v2 -- companion guide: https://zeroscope.replicate.dev/

---

For detailed usage, setup, and more, here is the official Replicate Documentation for Node.js.