Sessions & resume

Cloud jobs run on the server, so unlike a local export, they don't die with the tab. @eazip/core exposes just enough of the session to pick a job back up: the moment a cloud job's Public Session is created — not when the zip finishes — job.getSnapshot().session is populated:

const job = startZip({ strategy: 'cloud', publicKey, files });

job.getSnapshot().session;
// { sessionId, clientSecret, apiBaseUrl, createdAt, expiresAt, jobStatus, job }

Persist sessionId and clientSecret (session storage, a cookie, wherever survives your reload) as soon as they appear, and hand them to resumeZip after the reload to attach a fresh ZipJob to the same server-side job:

import { resumeZip } from '@eazip/core';

const job = resumeZip({ sessionId, clientSecret });
const result = await job.done; // polling picks up where it left off
result.download();

resumeZip doesn't care who created the session — the SDK itself, or a createSession callback backed by your own server (see Backend-created sessions) — so the same call works either way. It returns a ZipJob exactly like startZip, with one difference: there's no create step, so session is present in the snapshot immediately, before the first poll resolves it. Status is polling-only today — there's no webhook or SSE channel.

@eazip/react builds exactly this on top of localStorage — see Persistence & reload-resume if you're using the hook instead of @eazip/core directly.

Expiry

Sessions don't live forever. Resuming (or polling) a session past its expiresAt rejects job.done with EazipSessionExpiredError (SESSION_EXPIRED); a session an operator revoked rejects with EazipSessionRevokedError (SESSION_REVOKED). Both mean the same thing in practice — the credentials are no longer good — so re-run the export rather than trying to resume the old one.

Lower-level access: SessionsClient

For anything startZip/resumeZip don't cover — managing sessions outside the ZipJob lifecycle entirely — SessionsClient wraps the raw Public Sessions API: create, get, and poll, with the same backoff resumeZip uses internally.

import { SessionsClient } from '@eazip/core/cloud';

const client = new SessionsClient({ publicKey, apiBaseUrl: 'https://api.eazip.io' });
const created = await client.create({ files });
const session = await client.poll(created.id, { clientSecret: created.clientSecret });