Jobs

Every generation endpoint in this API returns a jobId immediately and keeps working in the background. This page covers the one endpoint you'll call repeatedly: checking a job's status and getting its result.

Get a job

GET /jobs/{jobId}

curl https://api.sloyd.ai/api/jobs/01JZQY4YJ7KV1GQZ9G6H6H2X25 \
  -H "x-client-id: $SLOYD_CLIENT_ID" \
  -H "x-client-secret: $SLOYD_CLIENT_SECRET"

Response

{
  "id": "01JZQY4YJ7KV1GQZ9G6H6H2X25",
  "status": "success",
  "title": "a medieval knight helmet",
  "prompt": "a medieval knight helmet",
  "createdAt": "2026-08-11T10:15:00.000Z",
  "genParams": { "targetFaceCount": 20000, "textureResolution": "auto", "topology": "auto" },
  "fragments": [{ "name": "i23d_queue", "displayName": "Waiting in Queue" }],
  "isSuccess": true,
  "libraryEligible": true
}

Field	Type	Notes
id	string	Same as the jobId you requested.
status	"pending" | "running" | "success" | "error"	Poll until this is success or error. Older jobs may show legacy values (created, completed, failed) with the same meaning.
title	string
prompt	string	The prompt used, where applicable.
genParams	object	The generation parameters that were applied, including defaults you didn't explicitly set.
fragments	array	Named processing stages with progress metadata — useful for a progress indicator, not required for correctness.
errorMessage	string	Present when status is error.
isSuccess	boolean	Mirrors status === "success".
createdAt	date-time

Fields not relevant to API integration (author display info, library/moderation metadata) are included in the response but can be ignored.

Downloading the result

Once status is success, the finished asset is available from a public, unauthenticated URL derived from the job id — there's no separate "download" endpoint to call:

https://storage.googleapis.com/ai-services-quality/jobs/{jobId}.glb

curl -o model.glb https://storage.googleapis.com/ai-services-quality/jobs/01JZQY4YJ7KV1GQZ9G6H6H2X25.glb

This applies to every 3D generation endpoint — each publishes its finished GLB at this same jobs/{jobId}.glb path, keyed on the jobId that endpoint returned.

TIP

Because the URL is public and predictable, you can start fetching it optimistically once you expect status to flip to success, or treat a 404/incomplete file as "not ready yet" instead of polling the JSON endpoint. Polling GET /jobs/{jobId} is still the reliable way to detect error.

No list endpoint (yet)

There is currently no supported endpoint to list every job created by your API key — track the jobIds you receive from each create call on your own side, and poll each one individually.