Batch image resizing and format conversion

Photoshop's APIs can process a document through a defined action or preset, such as resizing, cropping, or converting file formats across many images at once. A Beam agent triggers this when new assets land in a shared folder, reads each file's dimensions and format, applies the resize or conversion rule the team approved, and saves the output to the destination location under the agreed naming convention. If a file fails to process, arrives in an unsupported format, or the output looks visibly wrong, the agent flags it for a designer to review manually.

Layer and smart object updates

Photoshop documents are built from layers and smart objects that can be swapped or edited independently, such as replacing a background image or updating text in a template. A Beam agent detects when a new value is ready, for example an updated price or product photo, reads the target layer in the template file, and applies the swap through the approved script or action. It then exports the updated file for the requested channel. Any request to change layout, add a new layer type, or resolve a rendering conflict is routed to a designer instead of guessed.

Metadata and alt text generation

Photoshop stores file metadata such as keywords, captions, and alt text alongside the image data. A Beam agent can read a batch of newly uploaded images, apply the description rule a team has approved, for example pulling the product name and category into the alt text field, and write that metadata back to each file. It logs which files were updated and which were skipped because required fields were missing. Files with ambiguous content or no matching product record are sent to a person to describe and tag manually.