Guide
Asset Library FAQ
Group ownership, review duration, polling frequency, and image URL requirements
Asset Library FAQ
Common Questions
Groups created in the console are not visible in the API
Groups created in the console appear in GET /api/v1/materials/groups. If you only have a group ID, you can also use it directly as group_id in POST /api/v1/materials/assets; the platform first confirms that the group exists, then binds it to the current account, after which it appears in the group list.
Where can I find the reason for a review failure
The data.asset.error_message returned when querying the asset is the failure reason. It has content only when status is Failed; for normal and available statuses it is null.
How long does image review take
Review and preprocessing are asynchronous, and the time depends on the asset itself; the platform does not promise a uniform duration for all images. Your program should poll the same asset ID and must not resubmit just because the initial status is Processing.
Will frequent queries be rate-limited
Status queries are lightweight, but do not make repeated concurrent requests within a single polling cycle; it is recommended to query the same asset ID once every 5–10 seconds. When requests are too frequent, the API returns HTTP 429; simply retry later, and submitted assets are not affected.
What are the requirements for image URLs
The address must be an http(s) address that the platform server can download directly (both http and https are supported), and it must not depend on cookies, Referer, a login page, or an IP allowlist. If the image is behind an address that requires login, on an internal network, or available only as a local file, first transfer it to object storage or use data_url; otherwise asset processing will fail.
Whenever a public address can be provided, prefer image_url: the platform passes the address through as-is without downloading or transferring it, which is the fastest for review submission; data_url requires uploading the entire image as base64, is noticeably slower, and consumes platform storage.
