Neural Filters FAQ

Last updated on Oct 15, 2026

Get answers to common questions about Neural Filters requirements, downloads, and privacy.

These questions typically arise when exploring Neural Filters for the first time, troubleshooting download or performance issues, or understanding how Adobe handles your data. Questions are organized by setup and requirements, then privacy and inclusivity topics.

Neural Filters work on macOS 10.15 or later and Windows 10 or later. All Neural Filters are available on both operating systems, though some filters require an internet connection for cloud processing.

Most Neural Filters download to your computer and process locally. Some filters like Smart Portrait and Depth Blur (Beta) process some operations in the cloud and require an internet connection for full functionality.

Photoshop automatically notifies you when filter updates are available. Adobe continuously improves the machine learning models that power Neural Filters, and updates are delivered through the Neural Filters panel.

Filter download sizes range from under 1 MB for Skin Smoothing to over 650 MB for Style Transfers. Size varies based on the complexity of the underlying machine learning model, and Adobe works to optimize each model's size.

Neural Filters first try to use your computer's GPU for processing. If no performance GPU is available, filters automatically default to CPU processing. A high-performance GPU is not required but may improve processing speed.

Yes. Each subfolder contains one filter model.

  • On macOS, find filters at ~/Library/Application Support/Adobe/UXP/PluginsStorage/PHSP/23/Internal/com.adobe.nfp.gallery/PluginData.
  • On Windows, navigate to %appdata%\Roaming\Adobe\UXP\PluginsStorage\PHSP\23\Internal\com.adobe.nfp.gallery\PluginData.
  • Adobe collects anonymous behavioral data such as clicks, likes, and slider adjustments.
  • Adobe collects images anonymously without associating them with your identity when you submit feedback.
  • Adobe collects data to optimize machine learning models.

Adobe is committed to ensuring inclusivity and equity in Neural Filters. Unexpected outcomes may arise because filters frequently use new technology to process portraits. If you experience biased results, report them through the Neural Filters feedback mechanism so Adobe can address issues in real time.