eternium-sdk

Official JavaScript/TypeScript SDK for the Eternium API -- AI image, video, chat, embeddings, and audio generation.

Install

npm install eternium-sdk

Quick Start

import { Eternium } from 'eternium-sdk'

const client = new Eternium('etrn_your_key')

// Generate an image
const image = await client.image('A futuristic city at sunset')
console.log(image.url)

// Generate a video
const video = await client.video('Ocean waves on rocks', { model: 'veo-3', duration: 5 })
console.log(video.url)

// Run a multi-step pipeline
const pack = await client.runPipeline('social-media-pack', 'Product on white background')
console.log(pack.results)

Features

• 24 AI models (image, video, chat, embeddings, audio)
• Automatic polling until generation completes
• Built-in prompt caching (saves credits on repeated calls)
• Progress callbacks
• TypeScript types included
• ESM and CommonJS support

API

new Eternium(apiKey, options?)

| Option | Default | Description | |--------|---------|-------------| | baseUrl | https://api.eternium.ai | API base URL | | pollInterval | 3000 | Polling interval in ms | | timeout | 300000 | Max wait time in ms |

Methods

| Method | Description | |--------|-------------| | client.image(prompt, opts?) | Generate an image, wait for result | | client.video(prompt, opts?) | Generate a video, wait for result | | client.runPipeline(name, prompt, opts?) | Run a multi-step pipeline | | client.listModels() | List all available models | | client.listPipelines() | List available pipelines | | client.listTiers() | Get pricing tiers | | client.getUsage() | Check credit usage and budget | | client.getTaskStatus(taskId) | Poll task status | | client.getDownloadUrl(taskId) | Get download URL (20min expiry) |

Options

const result = await client.image('prompt', {
  model: 'nano-banana-2',     // Model to use
  aspect_ratio: '16:9',       // Aspect ratio
  cache: true,                 // Enable prompt caching (default: true)
  wait: true,                  // Wait for completion (default: true)
  onProgress: ({ status }) => console.log(status),
})

OpenAI-Compatible

The Eternium API also supports OpenAI-compatible endpoints. Use the standard OpenAI SDK:

import OpenAI from 'openai'

const openai = new OpenAI({
  apiKey: 'etrn_your_key',
  baseURL: 'https://api.eternium.ai/v1',
})

const chat = await openai.chat.completions.create({
  model: 'gpt-5.1',
  messages: [{ role: 'user', content: 'Hello!' }],
})

License

MIT - Eternium LLC