SDKs

ZeroGen publishes an official library for TypeScript and JavaScript, @zerogen/sdk. Its types come from the API’s OpenAPI document, and it follows the conventions in these docs for you:

It sends an Idempotency-Key with every POST, and reuses it when it retries.
It retries connection failures and errors whose retryable is true, and waits for Retry-After on a 429.
It waits for a generation to end with Prefer: wait.
It pages through lists, and uploads a file in one call.

The library runs on servers: Node.js 20.3 or later, Deno, Bun, and edge runtimes such as Cloudflare Workers. It has no dependencies. Its source is on GitHub. In other languages, call the REST API directly. The API reference and the OpenAPI document describe every endpoint.

Install

npm install @zerogen/sdk

Create a generation

Set your API key as ZEROGEN_API_KEY, then create a generation and wait for it to end:

import ZeroGen from "@zerogen/sdk";

const zerogen = new ZeroGen(); // Reads ZEROGEN_API_KEY.

const generation = await zerogen.generations.createAndWait({
  model: "openai/gpt-image-2/generate",
  prompt: "A lighthouse at dawn, in watercolor",
});

console.log(generation.status, generation.outputs);

createAndWait creates the generation with Prefer: wait=60, then reads it with the same wait until it ends. It resolves for every final status, including failed, so check status and each output’s status. To return without waiting, call generations.create, then generations.get or generations.waitFor with the generation’s ID. To stop waiting, pass an AbortSignal as signal; the generation keeps running. Each model takes its own inputs and parameters. Read a model’s request schema with zerogen.models.get("openai/gpt-image-2/generate"). For details, see Models.

Upload an input

uploads.upload runs the three requests in Upload a file and resolves with the ready asset’s ID:

import { openAsBlob } from "node:fs";

const upload = await zerogen.uploads.upload(await openAsBlob("storyboard.png"), {
  filename: "storyboard.png",
});

const edit = await zerogen.generations.createAndWait({
  model: "openai/gpt-image-2/edit",
  prompt: "Turn the sky into a sunset",
  inputs: [{ role: "source", asset: upload.assetId }],
});

It takes a Blob, File, ArrayBuffer, or Uint8Array, and reads the media type from the file or from the filename’s extension. For files up to 64 MiB, it sends the file’s SHA-256 digest, so a file the project already has is reused instead of uploaded again.

List

Iterate a list with for await to read every item on every page:

for await (const model of zerogen.models.list({ capability: "video" })) {
  console.log(model.id);
}

Awaiting a list instead returns one page, with data and nextCursor.

Handle errors

A request that ZeroGen refuses throws a ZeroGenError, which carries the error envelope: status, code, param, retryable, details, and the response’s requestId. Branch on code:

import ZeroGen, { ZeroGenError } from "@zerogen/sdk";

try {
  await zerogen.generations.create({
    model: "openai/gpt-image-2/generate",
    prompt: "A lighthouse at dawn, in watercolor",
  });
} catch (error) {
  if (error instanceof ZeroGenError && error.code === "insufficient_balance") {
    // Ask an admin to add funds.
  }
  throw error;
}

A request that gets no response, because the connection failed or timed out, throws a ZeroGenConnectionError.

Retries and idempotency

The library retries a request twice by default. It retries only connection failures and errors whose retryable is true, and it doesn’t wait for a Retry-After longer than one minute. Each POST call gets a new Idempotency-Key, which that call’s retries reuse. To retry safely across calls or processes, such as from a job queue, pass your own key:

await zerogen.generations.create(job.request, { idempotencyKey: job.id });

For details, see Idempotency.

Configure the client

Option	Default	Description
`apiKey`	`ZEROGEN_API_KEY`	The API key.
`baseURL`	`https://api.zerogen.com/v1`	The API’s base URL.
`maxRetries`	`2`	Retries for each request.
`timeout`	`120000`	Milliseconds before an attempt is abandoned. A request that waits for a generation gets its wait added.
`fetch`	The global `fetch`	A `fetch` implementation to use instead.
`defaultHeaders`	None	Headers to send with every request.

Every method also takes request options as its last argument: signal, timeout, maxRetries, idempotencyKey for a POST, and headers.

Get started

Generate

Conventions

Install

Create a generation

Upload an input

List

Handle errors

Retries and idempotency

Configure the client

​Install

​Create a generation

​Upload an input

​List

​Handle errors

​Retries and idempotency

​Configure the client

Install

Create a generation

Upload an input

List

Handle errors

Retries and idempotency

Configure the client