What's New?

2026-06-08 | File IDs Now Displayed for Partial File Download Errors in read_files and cache_files

Error messaging for read_files and cache_files has been updated to give you more information so you can quickly recover if some of your files don’t download successfully. Now you now see exactly which files failed and which ones you can keep using without rerunning your whole workflow.


• Understand what went wrong faster. When one or more files can’t be downloaded, the response now includes a message that might be helpful in determining why the download failed:

user’s role and file’s availability doesn’t permit access

  
  

• See exactly which files failed. Any files that don’t download are summarized in the output as Some files failed to download: [list of file IDs], along with an error message similar to the one shown above.

• Keep working with the files that did succeed. Files that download successfully in the same read_files and cache_files call remain available, so you can continue your analysis or processing while you address only the file IDs that failed.

To see the new behavior, submit a list of file IDs to read_files or cache_files as you normally would. If any files fail to download, the response includes an error message plus the “some files failed to download” warning and the list of affected file IDs.

2026-06-08 | Visualization Support for Pixi Builds
HISE now offers Pixi support for visualizations, making it possible to build an app from your IDE using an existing Pixi environment. You can call save_visualization_app() directly from your IDE, use a new Pixi-powered build template, and rely on Pixi across the build and run stages.

• Unified dependency definitions. The save_visualization_app() SDK method now works seamlessly from Pixi-based IDEs, sending a Pixi-compatible environment file into the build process so you don't have to maintain separate dependency definitions.

• Faster visualization builds.The updated visualization build path now uses Pixi as the package and environment manager. It also introduces a template labeled Dash with Pixi, without pre-installed packages (v3.0), so that build uses your Pixi environment instead of preinstalled packages.

• Visualizations aligned with your Pixi workflow. Moving your visualization builds to Pixi means the build pipeline uses the same Pixi environment you already rely on in your IDEs. You define dependencies only once. The deployed visualization behaves the same way it did in development. No surprises.

To use this feature, create or open a Pixi-based IDE and call save_visualization_app(). When prompted, select the template labeled Dash with Pixi, without pre-installed packages (v3.0). For details, see Save Visualizations in HISE (Tutorial). If you have questions or need help, contact Support.

2026-05-26 | AI/ML Stacking Helps You Organize Related Environments

You can now stack related AI/ML environments and training runs into ordered groups to highlight active layers and show the relationships among the layers, while keeping older iterations for context or comparison. Each layer is labeled by name, description (optional), modified date, and status (such as Available). As your research evolves, you can move, remove, add, or delete layers or dissolve stacks to accommodate your workflow. The underlying environments and their reproducibility are preserved.

• Try it yourself. Go to the AI/ML hub (Research > AI/ML) to see the new stacking view. To try it, create a new stack from two or more environments or training runs.

• Reorganize layers as your analysis evolves. Stacking gives you the flexibility to add new layers, move layers, or set a new training run or environment at the top of the stack.

• See only what you need. You can show or hide deleted environments and training runs, remove layers, or dissolve a stack while keeping all underlying components intact for reproducibility.

For details and step-by-step instructions, see Create an AI/ML Stack (Tutorial). If you have questions or need help, contact Support.

2026-05-14 | Private Folder Access from Your Laptop

You can now reach your existing HISE private folder directly from your laptop. This update lets you use gsutil to move files between HISE and your laptop as you explore data. To make it easier to upload files from your laptop, HISE also now displays your private folder’s cloud storage bucket name at the new Private Folders link on the main menu (coming soon!).

For details, see Get Direct Access to Your Private Folder (Tutorial) and Download Files to Your Laptop (Tutorial) .

If you have questions or need help, contact Support .

NOTE
You can use gsutil to copy files or folders directly from your laptop into your HISE private folder. Be aware that using this workflow can compromise reproducibility. In addition, data stored in private folders counts toward your team's cloud billing quota, so be sure to remove or delete files when you no longer need them.