
An incoming Cloudinary transformation changes the asset before it is stored. An eager transformation creates an additional derived rendition. If your purpose is to keep a full-resolution master and prepare a thumbnail, put the thumbnail operation in the derivative stage and verify the stored master independently.
Key Takeaways
- Incoming processing can intentionally change the stored source.
- Eager processing prepares additional delivery renditions.
- A thumbnail preview does not establish that the master was preserved.
Find where the operation was configured
Cloudinary's upload-parameter guide describes incoming transformations through the upload transformation parameter, applied before storage. It separately describes eager transformations that generate additional transformed assets.
Inspect the actual upload request and the selected upload preset. A transformation may originate in the preset even if the browser request looks harmless. Preserve the local source while diagnosing; do not overwrite your only full-resolution copy.
Choose the storage contract first
- Write down the required stored master dimensions and the separate thumbnail requirements.
- Inspect request and preset definitions for incoming transformations.
- Determine whether reduced dimensions, a crop or another baked-in change is intentional.
- If the master must remain unchanged, have the authorized owner prepare a sample configuration that creates the thumbnail eagerly instead.
- Upload a disposable sample and inspect the stored source's dimensions and appearance.
- Retrieve the intended thumbnail separately and compare it with the approved crop.
A successful upload response only confirms what the service accepted under that configuration. It cannot prove that a pre-storage transformation preserved pixels outside a crop.
A hypothetical product-photo handoff
A studio keeps an approved 4000-pixel photograph locally and wants a 400-pixel shop tile. These dimensions are a planning example. An incoming resize can make the stored copy smaller; an eager tile is meant to coexist with the uploaded asset.
The proposed acceptance record lists two files: the stored source and the tile. Check each one's actual dimensions, crop, transparency where relevant and visible product details. Do not compare only the tile because that is the file the storefront currently displays.
Keep provider terminology explicit
Official ImageKit pre/post transformation guide documents the analogous distinction: pre-transformation occurs before upload, while post-transformations are generated in addition to the original. That comparison helps translate the storage decision; it does not mean the two providers share request syntax or processing limits.
The master-versus-flat-output guide addresses editing structure. This issue is different: which pixels are retained as the stored source at ingestion. If an incoming crop already discarded content from the stored asset, changing a delivery URL cannot restore it; return to an independently retained original.
Related tasks
- What Is the Difference Between an Editable Artwork Master and a Flat PNG?
- How Do I Keep the Approved AI Image or Video from Getting Lost Among Revisions?
Sources and scope
- Cloudinary: Upload parameters: Eager transformations add derived renditions; incoming transformations run before storage; eager_async with eager_notification_url enables background processing.
- ImageKit: Pre and post upload transformations: Pre-transformations alter media before upload; post-transformations are created in addition to the original; provider-specific request syntax and asynchronous behavior.
Documentation-based editorial guidance. All scenarios, worksheets and acceptance checks are hypothetical or proposed. No hands-on test, account operation, benchmark, measured search demand, AI-citation rate or platform acceptance is claimed.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: AI-generated conceptual editorial illustration; not a real application screenshot, observed output or before-and-after test.