Here's what's new in our latest update:

Faster Results Loading with Local Render Data Caching

To improve performance, Luminary now caches visualization data files locally using your browser. This feature substantially improves project load times (up to 45x in some cases) when re-opening previously visited projects, as the data no longer needs to be downloaded. You can manage your cache size in My Account > Settings. The cached data is managed securely by your browser and not visible to other pages or your local filesystem.

AI Assistant: Expanded Output Support

The Assistant now supports creating, updating, and deleting a wider range of outputs — including force/moment, surface average, residual, and volume reduction. You can also modify reference values and stopping conditions directly through the assistant. (Note: Point probes, inner iteration data, and derived outputs aren't yet supported.)
​

Enhancements

• Blender versions up to 4.5.1 LTS are now supported
• You will now be prompted to confirm before the AI Assistant generates a mesh
• Free trial users now have access to the in-app notebook
• When you rename a simulation, any final adapted mesh it generated will now update its name to match, keeping things consistent and ensuring your project stays organized without manual cleanup

Fixed Issues

• The notebook crashed in some instances
• Camera grouping in geometry did not function properly in some cases

Here's what's new in our latest update:

Collaborate on Projects with Writer Access Sharing

You can now share projects with Writer permissions. This allows multiple users within your organization to collaboratively edit projects in real time.

In-App Notebook: Explain Code

You can now ask the assistant to explain code directly from the hover toolbar within a notebook cell. Just hover and click Explain Code to get an instant breakdown of what’s happening.

Convert Colors to Tags

You can now convert surface colors from your CAD file into Tags. This gives you a visual and flexible way to organize surfaces before importing into Luminary. To use it, right-click on Tags in the geometry tree and select Convert Colors to Tags.

In Case You Missed It: Luminary's MCP Server

The new Luminary Cloud MCP Server lets your AI agents in VSCode and Cursor connect directly to the Luminary platform. This enables them to write Luminary SDK code within your IDE.

Learn how to install the Luminary MCP server here.
​

Multi-Kernel History
You can now switch between continuous and discrete kernels without breaking model history. This unlocks a more fluid and intuitive workflow across geometry types.

• Automatic Conversion: Shrinkwrap and discrete-part imports now auto-convert geometry on the fly.
• No Manual Conversion Needed: The old Convert to Discrete button is now obsolete.
• History-Safe: Rolling back operations automatically restores the original continuous state, maintaining full history integrity.

This enhancement eliminates friction when transitioning between modeling and meshing workflows, while preserving traceability and flexibility.

Enhancements

• emoved redundant warnings and checks related to rotation angles and translation displacements
• Free trial users can now access and use the Luminary SDK.
• Mesh generation by the AI Assistant now requires user confirmation, helping prevent unexpected actions.

Fixed Issues

• Subtracting without deleting tool bodies no longer deletes their tags.
• Updated the minimum quantity requirement in geometry patterns to enforce positive values and correctly handle body deletion based on the pattern’s direction and quantity.
• Added validation to require at least one surface when using Selected Surfaces mode for both meshing and modeling.

Here's what's new in our latest update:

Supercharge Your Development with AI That Actually Knows Luminary Cloud

Your AI coding assistant just learned our entire SDK. The new Luminary Cloud MCP Server connects your AI agents in VSCode and Cursor directly to our platform, giving them insider knowledge and live project access.

What You Get

Faster development. Your AI understands our APIs, best practices, and your specific projects—no more generic suggestions or outdated examples.

Less debugging. Get exact patterns and functions for your simulation pipelines instead of hunting through docs.

Zero friction. Works instantly with VSCode, Cursor, Claude, and Gemini CLI.

Key Features

• Deep SDK knowledge with every function and best practice
• Live access to your actual Luminary Cloud projects
• Context-aware code generation for simulation scripts
• Real-time help without leaving your IDE

Get Started in Minutes

Ready to experience AI-powered development with Luminary Cloud?

Set up your MCP Server →

Installation takes less than 5 minutes with our one-click setup for both VSCode and Cursor. Once connected, just ask your AI assistant about your Luminary Cloud projects and watch the magic happen.

Here's what's new in our latest update:

In-App Notebook

We’ve launched an AI-assisted, zero-setup Python Notebook to streamline scripting, visualization, and post-processing with the Luminary SDK—integrated directly into the application. Skip environment setup and SDK installs; just open the notebook, and you're ready to start coding.

The integrated AI assistant helps generate and complete scripts quickly, while an interactive 3D widget lets you explore simulation results, compare runs, and analyze data without leaving within the platform.

You can access the notebook from the sidebar in the Projects page. Read more about its features and best practices here.

Enhancements

• When the assistant analyzes data, the generated code will now be provided so you can use it in your own scripts

Fixed Issues

• General performance improvements and bug fixes

Here's what's new in our latest update:

Enhancements

• Iterations per output now defaults to zero. For LMA runs, this will result in a solution being written out after each adaptation run.

Fixed Issues

• Resolved an issue with internal fan curves.

Here's what's new in our latest update:

Enhancements

• Robust startup is now the default for default solution controls
• General performance enhancements

Fixed Issues

• Resolved an issue when defining a refinement region via the SDK

Here's what's new in our latest update:

Enhancements

• Failed LMA meshes can now be used for simulations, allowing you to test different settings on the mesh
• Volume download is now available for failed and suspended simulations
• You can now hide all tags in the geometry tree

Fixed Issues

• General performance improvements

Here's what's new in our latest update:

Enhancements

• You can now hide all Tags in the Geometry tree
• Use the SDK to extract Intersection Curves

Fixed Issues

• Resolved an issue with LMA mesh interpolation

Here's what's new in our latest update:

Select All in Current View

You can now click Select All from the contextual right-click menu to select everything in the current view, excluding hidden surfaces.

Results Table: Download All

The Download All button in the Results table now includes additional columns: credits, run time, iterations, date, and run status.

Enhancements

• You can now plot the point coordinates of an intersection curve or line plot
• Right-clicking on a tag now provides all the same options available for surfaces—such as Hide and Hide Others (Isolate)
• The visualizer now maintains the center of rotation, improving zoom-to-cursor functionality and directional surface inspection
• We've added data analysis commands to the AI Assistant's suggested items

Fixed Issues

• Applying settings from the library no longer removes the job priority setting
• The caret icon has been restored for the project dropdown menu. As before, the menu now opens on left-click
• Some monitor points in transient simulations were not updating—this has been fixed.
• Taking screenshots in the Setup tab now works as expected

Here's what's new in our latest update:

Reference Frame for Spatially Varying Heat Profile BC and Porous Media

• For spatially varying heat boundary conditions, you can now view the reference frame associated with each surface, which will be used to apply the spatial variation.
• For porous media behaviors, you can now view the reference frame associated with each volume, which will be used to apply Darcy and Forchheimer coefficients.

Note: This feature is currently incompatible with tags.

Tags as References for Meshing

Tags can now be used as references to volumes for mesh sizing, supporting robust mesh automation pipelines.

OpenCSM Integration

We now support integration with OpenCSM, developed by Engineering Sketch Pad. OpenCSM is an open-source CAD package that enables parametric geometry modeling.

Enhancements

• The project dropdown menu (previously accessed by clicking the project name in the header) has been updated to a right-click context menu
• You can now define the Value Range for Threshold Plots using "Less than" and "Greater than" conditions
• Support has been added for newer versions of many CAD file formats
• You can now plot X/Y/Z coordinates on 2D plots for both intersection curves and lines, when using client-side rendering
• SDK release 0.15.1 is now available, introducing support for cloning a Scene

Fixed Issues

• Resolved an issue preventing inversion of heat flux using a negative value