everpod

JavaScript client for Everpod — GPU inference on Household AI Infrastructure. Zero dependencies; Node 18+, Deno, Bun, and browsers.

npm install everpod

import { Everpod } from "everpod";

const client = new Everpod({ apiKey: "evp_..." }); // or EVERPOD_API_KEY

const image = await client.images.generate({ prompt: "a lighthouse at dusk, cinematic" });
console.log(image.url);

Generation runs as a job on a real GPU node; generate() polls to completion for you. For async control:

const job = await client.images.submit({ prompt: "a lighthouse at dusk" });
const done = await client.jobs.wait(job.id, { timeoutMs: 120_000 });
console.log(done.artifacts[0].url);

Check live capacity before submitting — models are only offered while a node can serve them:

for (const model of await client.models.available()) {
  console.log(model.name, model.readyWorkers);
}

Get an API key at app.everpod.io/keys · API reference at api.everpod.io/docs.