API 레퍼런스 · Media
Image · job status
Poll an async render until the image is ready.
Free to call, as often as you like — every few seconds is the intended cadence. status walks processing → succeeded (the image is in data) or failed (an error object explains, and the credits were already refunded).
While processing, phase tells you where the render is: queued (waiting for a free render slot — a busy period can hold it there for a while) or generating (a slot picked it up and the frames are being drawn). Both are healthy — keep polling.
A finished job carries generation_ms — the wall-clock the render took end-to-end (queue + generation), in milliseconds. Handy to show your own users how long a render ran.
Job ids are scoped to your account (and workspace) and expire 24 hours after creation; persist the image (yours to keep, or store: true at submit) before then.
data[0].url is a signed CDN URL valid for 30 minutes — poll the job again (or read /v1/creations) for a fresh one. A private render (private: true) answers once with b64_json instead and is stored nowhere.
curl https://eroq.ai/v1/images/generations/GENERATION_ID \
-H "Authorization: Bearer $EROQ_API_KEY"응답
{
"id": "b7e6c2d4-…",
"object": "image.generation",
"status": "succeeded",
"created": 1756118400,
"model": "eroq-krea2",
"data": [{ "url": "https://cdn.eroq.ai/…/image.webp?token=…&expires=…", "content_type": "image/webp", "library_id": "…" }],
"generation_ms": 4200,
"usage": { "credits_spent": 10, "credits_remaining": 977 }
}