Release v25.9.1
Neurotorch version 25.9.1 (released 04.09.2025)
This update is the biggest update yet. Started in March 2025, the codebase was completly rewritte for stability and expandability taking much longer than expected. Due to the nature of this core update, no in between versions were released. Therefore it is not possible to list all the changes completly, but the most important features are:
- Added support for MacOS and Linux (experimental): The code has been checked to support MacOS and Linux. However, this feature is not yet tested on those platforms (!)
- Faster loading: The core functions have been updated to support advanced and faster loading
- Logging system: Neurotorch is now equipped with an logging system helping to find bugs much earlier
- AppData: All config files, user plugins and the environment are now located in the users AppData folder
- New plugin system: The plugin system has been completly rewritten. It now supports an easy to use event API, better crash handling and support inside the GUI
- Session system: Neurotorch has been updated to a session based system. This allows to have multiple instances running in parallel (only one GUI) and may allow in the future to have multiple files open
- Filters: The new filter system replaces the old convolution functions. Filters allow more nuance modification of the delta video and can easily be stacked upon. Plugins may easily add new filters
- GUI: The GUI has been redesigned regarding user feedback
- Documentation: The documentation is now build from inside the code to allow more efficient updates
- Portable and standalone versions: Starting with this release, a portable, standalone version for Windows, MacOS and Linux are released on GitHub using Actions. Also, new releases are now directly updated from inside GitHub actions to PyPI
- Exporting: Neurotorch now allows to export videos, for example without the stimulation
- Settings: Improved settings API
- Task system: Improved task system
- and many more changes