MaplePrimes Announcement

Every engineer knows the part of the job that never makes it into the design review highlights. It’s the worst-case check you redo because a part changed. It’s the sizing calculation rebuilt for the third project in a row. It’s the report that has to record every step, so someone else can follow your work and sign it off.

Today we’re launching Maple Loop, an AI engineering agent from Maplesoft built to take on exactly that work. Here’s what it does.

Give it a problem, not a formula

Working with Maple Loop starts the way you’d brief a colleague. You describe the problem and point it to your files: schematics, datasheets, requirements, test data. Here’s a prompt from one of our worst-case circuit analysis runs:

“Our 3.3 V rail monitor resets the processor when the rail drops. The reset has to happen between 3.00 V and 3.20 V. Will that work with the parts we chose?"

From there, Maple Loop plans the work and moves through it the way an engineering team would: requirements, design, analysis, verification. It checks each phase before moving on. When it’s done, it comes back with the tolerance build-up for each part, the part costing the most margin, and a ranked list of fixes with what each one buys you.

You can start from a blank prompt like this one, or from a template for common jobs such as worst-case circuit analysis or PID control design. See it in action. 

Why the math runs in Maple

This is the part we care about most. General AI tools are good at producing answers, but in engineering an answer you can’t verify isn’t worth much. A plausible number from a language model can be wrong in ways that are hard to spot, and it gives your reviewers nothing to check.

So, Maple Loop splits the job.

 

You get the engineering, not just the answer

What comes back isn’t a chat transcript. It’s working files in tools you already know, with the full calculation inside.

Calculations arrive as Maple Flow documents, laid out like a clean hand calc with units carried through every step, ready for review. Full derivations come as Maple worksheets. And when a problem calls for system behavior, Maple Loop builds a MapleSim model you can simulate and tune.

The rest of the team doesn’t need Maple to benefit. From the same checked calculation, Maple Loop can produce Word reports, PDFs, Excel workbooks, PowerPoint decks, generated code and shareable dashboards.

Because everything is built on live Maple documents, the work doesn’t go stale. When a part changes, you change the input and rerun, and the analysis and documents update together. The next design revision starts from a working calculation instead of a blank page.

A new way to put Maple to work

Maple has always been about getting the math right. With Maple Loop, we’re putting the same engine to work in a new way, at the center of an AI agent that carries engineering projects from the first prompt to a reviewed result. It’s a natural next step for Maple: the same math you already rely on, now doing the work inside an engineering workflow.

Get access to Maple Loop

We’ve been testing Maple Loop with engineering teams on real problems across electrical and mechanical work, and many more.

Join the waitlist for access to Maple Loop.

What routine analysis would you hand off first? Which workflows should we focus on next?

Tell us in the comments.

Featured Post

Can the Maple community make exact integrability computations substantially faster?

For a planar polynomial vector field
[
D=A(x,y)\partial_x+B(x,y)\partial_y,
]
a Darboux polynomial is a nonconstant polynomial (f) satisfying
[
D(f)=Kf
]
for a polynomial cofactor (K). These invariant algebraic curves are fundamental building blocks in the search for first integrals and integrating factors.

The subject reaches back to Darboux’s nineteenth-century work and connects with the Prelle–Singer and Singer theories of elementary and Liouvillian integrability. Turning these structural results into effective, reliable programs remains an important computational problem. Given Maple’s long tradition in symbolic differential equations, MaplePrimes seems an ideal place to invite a practical challenge.

The starting point

Bostan, Chèze, Cluzeau and Weil developed algorithms for bounded-degree rational first integrals and Darboux polynomials using local solutions and extactic constructions. Related ideas appear in Chèze–Combot’s algorithms for symbolic first integrals and Combot’s work on symbolic integration of planar differential foliations. These works include Maple implementations.

They provide powerful mathematical frameworks, but their implementations can encounter substantial practical difficulties:

  • coefficient growth in exact truncated series;

  • expression swell in parameter-dependent extactic matrices, polynomial minors and Smith-form computations;

  • expensive matrix assembly, exact linear algebra and reconstruction.

The bottleneck depends on the input and the implementation. Improving the kernel solver alone may achieve little when exact series construction dominates.

Our current baseline

We have implemented practical optimizations within these frameworks: jet-based constructions, truncated-series arithmetic, avoidance of full Smith-form computation in the Darboux code, modular filtering and linear algebra, rational reconstruction, and exact verification.

These are our current best-performing implementations, offered as reproducible baselines for further improvement—not as a claim of universal optimality.

For Darboux polynomials:

restart:
read "PATH/TO/FastDarboux.mpl":

A := 1:
B := 4*x^3 + y - x^4 - 1:
N := 4:

R := FastDarboux:-All(A,B,N,x,y):
R[status];
R[curves];
R[rational_curves];

The records in rational_curves are pairs [f,K]. Algebraic coefficient extensions are represented in curves. The status "finite_complete" certifies complete finite enumeration within the stated scope; rational-first-integral and other exceptional cases have separate statuses.

A single-section calculation is available through:

R := FastDarboux:-Partial(A,B,N,0,x,y):

This searches using the section (x=0); its scope must not be confused with complete enumeration.

For the Liouvillian integrating-factor benchmark:

restart:
read "PATH/TO/Combot.mpl":
read "PATH/TO/slow_example.mpl":

st := time[real]():
R := IntegratingFactor(B/A,[-1,0],27):
time[real]() - st;

The benchmark file supplies (A,B). The implementation prints a completion status.

For this example, on our Maple 2026 test system, the complete call took about 123 seconds:

Exact series construction 100.4 s
Matrix preparation, modular kernels, reconstruction and kernel certification 21.9 s
Total, including integrating-factor recovery 123.1 s

The returned integrating factor was independently verified by exact identities. These measurements describe one benchmark, not a general complexity claim.

Source code: Source code-Huang 
Documentation and mathematical scope: Documentation-Huang 
Benchmark inputs, environment and logs: Benchmark-Huang

The challenge

Improve end-to-end wall-clock time and memory use while preserving soundness and completeness for an explicitly stated bounded problem.

A better asymptotic bound is welcome, but it is not the primary objective. In exact symbolic computation, a field operation has no fixed machine cost: coefficient sizes, normalization, intermediate representations and data movement can outweigh an attractive exponent.

Technical improvements and well-designed heuristics are especially welcome. Possible directions include finite-field series construction, better precision management, adaptive reconstruction, structured linear algebra, improved candidate selection, or entirely different approaches.

Heuristics are welcome when their role is explicit: a failed heuristic must not silently become a proof of nonexistence. Likewise, verifying every returned curve proves soundness, but does not by itself prove completeness.

Please state degree bounds and assumptions, distinguish “no solution within the bound” from “inconclusive,” and handle rational-first-integral or infinite-family cases explicitly. This challenge does not assume that bounded searches settle unrestricted Liouvillian integrability. The current integrating-factor code also does not claim to return a minimal representation.

For performance comparisons, please include the whole computation, including preprocessing, series, reconstruction and certification, and compare tasks with the same mathematical scope.

AI tools have made remarkable progress in mathematical reasoning and programming. This is a concrete opportunity to use them on a longstanding computational problem. Contributions from anyone are welcome, with or without AI assistance: Maple code, compiled backends, mathematical arguments, counterexamples, profiling observations, or small improvements that make an important difference.

Please post suggestions, implementations, benchmarks and supporting arguments directly in the comments. The goal is a faster, reproducible and mathematically trustworthy solver.

Featured Post

Ask any engineer where the time goes in a modeling project, and the answer is rarely "the math."

It's everything around it: hunting for commands, opening help in another window, checking a plot, finding your place in the model again. Each interruption is small. Together, they can turn an afternoon of modeling into a week.

MapleSim 2026 was built around that idea. From the first install to the final result, it keeps what you need right where you're working.

Getting started is now simpler than ever. MapleSim, all its toolboxes and the Maple engine are now delivered in a single installer, so one installation provides everything needed to get started.

Context-sensitive hover bars display the most frequently used commands next to the current selection, and consolidated menus group related commands together. These changes reduce the steps required to build and modify models.

A new split pane, which can be arranged beside the diagram or opened in a separate window, updates with each selection to display the component's help page, its Modelica code, or live plots of its variables in the time or frequency domain. Engineers can review component details without leaving the model.

The Optimization app now displays live system plots and a time history of the objective function during execution, allowing engineers to assess convergence and intervene early when needed. The in-product apps have also been redesigned with a more compact layout for use alongside the split pane.

FMI import and export now support string and discrete parameters, inputs, and outputs. The MapleSim CAD Toolbox supports the latest formats from Catia® V5, SOLIDWORKS®, Creo®, NX™, Inventor®, Solid Edge®, Parasolid®, and other CAD tools. MapleSim 2026 also supports the latest Maple interface, including ribbons, the AI Assistant, and improved performance.

Building a web line model involves many components, each with its own connections and parameters. The new Web Handling Model Creation App automates this process. Starting from a tabulated data file describing the line, the app validates the data, calculates wrap angles, span lengths, and roller mass and inertia, and generates the complete MapleSim model. Engineers can then proceed directly to adding drives, dancers, and control loops to analyze line performance.

New components extend modeling coverage to more of the processes found on real lines:

  • Noncircular winding and unwinding, for rolls that are not perfectly round

  • Delamination, where one web separates into two

  • Web-on-web friction, for two webs running through a nip roller together

  • Tendency drives, using the new Roller Shaft component

The library is also easier to use and delivers more accurate results. A reorganized Rollers sub-palette groups components into Rollers, Nip Rollers, Multi-Web, and Parameters. Sequential and S-Wrap Nip Rollers simplify S-wrap configurations, and a per-unit option adds flexibility to drives and controls. Improved handling of span tension loss and special nip roller configurations further increases accuracy.

MapleSim 2026 adds new components that capture more real-world detail in machine models.

  • Ropes and Pulleys Library: Variable Winch 3D allows engineers to define the shape of the winch drum. Pulley Edge models side contact between the sheave and the rope, and Pulley Multi-Turn models multiple rope turns for added friction, complete with visualization.

  • Multibody Library: A new Gear Forces sub-package provides simplified models for spur, helical, bevel, and crossed gear pairs, along with a ratchet wheel model.

  • Directional Control Valve app: The app now creates pneumatic valves in addition to hydraulic valves.

MapleSim 2026 reduces the time spent navigating the software and increases the time available for what matters most: understanding and improving designs.

To try MapleSim 2026, start your free instant evaluation today. For a complete overview of the release, visit What's New in MapleSim 2026.