
For Cloudinary PSD delivery, pg_1 represents all layers, not the first isolated layer. The documented bottom layer is pg_2. Select a layer by the correct index or its case-sensitive name, then inspect the returned image; a page-looking number is not sufficient evidence of which PSD content was requested.
Key Takeaways
- PSD pg indexing includes a whole-image entry.
- Named layer selection is case-sensitive and duplicate names need disambiguation.
- Original PSD delivery can expose the complete layered file.
Establish whether the request needs a composite or a layer
Keep the source PSD and identify the element the recipient needs. A finished poster, a single illustration and an editable master are different outputs. The master-versus-flat guide helps state that contract before selecting a delivery parameter.
The Cloudinary layered-media reference explicitly maps pg_1 to all layers and pg_2 to the bottom layer. Do not copy a PDF page interpretation into a PSD workflow just because both use pg.
For a hypothetical poster containing a backdrop, portrait and title, make a private layer map with recognizable content. The map is a proposed diagnostic aid, not a claim that those layers were uploaded or rendered.
Choose a documented layer reference
Use the correct documented layer index for the intended source. Alternatively, reference its case-sensitive name through pg_name. The official guide says that repeated names can be disambiguated with indexed name forms: the first instance uses the name, and the next uses name-1.
Check each selected result against its visible content before composing a multi-layer delivery. Cloudinary says selected layers are delivered in the specified order, which can differ from the original PSD ordering. A syntactically accepted list is therefore not approval of the resulting stack.
Do not assume that every layer and effect has an independently addressable image. The documentation warns that some effect or blend-mode layers cannot be accessed with pg.
Review mask behavior and original-file exposure
The same reference says PSD clipping masks work when the image is delivered as a whole, but not when delivering selected layers. Layer masks have different documented support. Inspect the intended transparency and edge rather than treating a missing clipped region as an index typo automatically.
Also inspect the actual output format. Delivering the original layered PSD or TIFF allows users to download its layers. A flat preview beside that link does not conceal the original. The hidden-artwork handoff guide explains why visible appearance alone cannot certify what a native file contains.
Accept identity, rendering and access separately
For a chosen-layer preview, confirm the correct subject, mask boundary, stack order and final image format. For a layered-source handoff, confirm that the complete source is intentionally included.
If the request instead concerns the original content of an embedded Smart Object, follow the embedded-object export guide or the applicable documented Cloudinary embedded-object option. Do not describe a selected-layer raster as that original by default. No layer extraction or access-control test is reported here.
Related tasks
- Export an Embedded Smart Object Without Exporting the Whole Poster
- What Is the Difference Between an Editable Artwork Master and a Flat PNG?
- Does an Illustrator Clipping Mask Actually Remove the Artwork Outside the Crop?
Sources and scope
- Cloudinary: Paged and layered media: PDF image delivery defaults to the first page and supports pg page selection and dn raster density, whose default is 150 DPI. For PSD delivery, pg_1 represents all layers and pg_2 the bottom layer; names are case-sensitive. Original layered files retain their layers, and incoming transformations can discard source layer data.
Documentation-based guidance from the cited official pages captured on October 25, 2026. Scenarios and acceptance procedures are proposed, not executed tests. No account operation, measured demand, AI-citation frequency, processing benchmark or search-indexing outcome 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.