Start one job or a batch
Needs the scope jobs:write
Start one job, or up to 50 as { jobs: [...] }. Jobs run our production services on a delivered product: packshot renders, AI scenes, image edits, videos and 3D exports. Every job type runs.
- Jobs queue, and the results are added to the product: new images appear in images[] with the job's id, asset.images.updated is sent, and job.succeeded carries the job with its outputs.
- Pictures are made in the order asked, and each is saved under its fileName from the start.
- A batch answers a results array like POST /api/v1/products.
- A product without a delivered model answers 409 not_ready (edit jobs work on an image and need no model).
- AI jobs (scenes, edits, scene videos) run for an account with an allowance of their kind; without one they answer 403 forbidden.
- An export to a format not switched on for your account answers 403 forbidden.
Idempotency-Key?stringA unique value such as a UUID. A retry with the same key within 24 hours returns the original answer and runs nothing twice; the same key with a different body is refused with 409 idempotency_conflict.
length <= 128application/json- body
type*stringpackshot: studio images from the model. scene: AI lifestyle images with the product. edit: a change to an image. video: a turntable or a scene video. export: the model in other file formats.
"packshot""scene""edit""video""export"articleId*stringThe product to work on. It must be delivered, except for edit jobs, which work on an image.
variant?stringYour variant code, for models with several material variants.
length <= 64options?||||The options of the job type. Anything left out takes the default.
naming?A naming template for this job only.
metadata?Up to 20 key and value pairs of your own, such as an order number. Returned on the job and in its events.
Accepted
application/json- response
id*stringuuidtype*string"packshot""scene""edit""video""export"articleId*stringvariant*string|nullstatus*string"queued""running""succeeded""failed""cancelled"options*The options as applied, defaults filled in.
naming*|nullmetadata*createdAt*stringdate-timestartedAt*|nullfinishedAt*|nulloutputs*array<>The files made, in the order asked for. Empty until the job finishes; a failed job lists the files it did make.
error*|Why the job failed, when it did.
curl -sS -X POST https://platform.charpstar.co/api/v1/jobs \ -H "Authorization: Bearer $CHARPSTAR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "type": "export", "articleId": "CH-1001", "options": { "formats": ["fbx", "obj", "glb"] } }'{ "id": "5c0e2f9a-7b14-4d3e-a8c1-93f6e2b7d045", "type": "packshot", "articleId": "CH-1001", "variant": null, "status": "queued", "options": { "views": [ "front", "left", "back" ], "background": "transparent", "size": 4096, "aspectRatio": "1:1", "format": "png", "shadows": true, "lighting": "studio", "preset": "Marketplace (GS1)" }, "naming": { "preset": "gs1" }, "metadata": { "purchaseOrder": "PO-88213" }, "createdAt": "2026-11-02T08:00:00.000Z", "startedAt": null, "finishedAt": null, "outputs": [], "error": null}Read many articles, or what changed GET
Pass articles (up to 100, comma separated) to read several at once, or since to get every article that changed after that time, oldest change first, 100 a page. With since, keep calling with cursor set to nextCursor until it is null. A product can appear in two runs in a row; write its current state, keyed on articleId.
List jobs GET
Your jobs, newest first, with their status and outputs. Filter by article, type, status, or changes since a time.