StarLogo Nova is an agent-based game and simulation programming environment that combines an easy-to-use blocks-based programming language with a powerful simulation engine and 3D renderer. With StarLogo Nova you can:

Create, edit, and run games and simulations right in the browser, no installation necessary.
Share projects in public galleries for the world to see.
Collaborate on projects with other users.
Incorporate your own sounds and 3D models into your projects.
Organize code more clearly, with all runtime code now placed on breed pages.
Program agent interactions more easily with new Detection blocks.
Customize your breeds with user-created traits like energy, health, lives, inventory, etc.
Easily work with hundreds of agents, even on older computers or Chromebooks.
With no predefined agent limits, create 10,000 agents or more on powerful computers.

MOST RECENT UPDATE

November 2, 2017
We've just released an update with the following improvements:

• Some users were encountering frequent WebGL crashes, which were crashing the entire computer for some versions of Mac OS. This issue has been fixed for most users, and performance has been improved for large numbers of agents. If you still encounter WebGL crashes, please contact us using the link at the bottom of the page.
• Renaming breeds previously left some blocks attempting to use the old breed name. This has been fixed.
• Copying, pasting, saving, and loading of procedure blocks and procedure call blocks has been overhauled to fix a variety of issues related to call blocks not remembering which procedure they were calling.
• The "return early" block now updates itself to match the return type of whatever procedure you connect it to.
• Old links to galleries correctly open that same gallery in the new interface.
• Links to public galleries work even if you are not signed in.
• Fixed a bug where sliders could not have a minimum or maximum set to 0.
• The Drawer Selector has been re-styled and shows colors properly on all browsers.
• Fixed a bug where the "nearest" block was not actually finding the nearest agents.