Over the past year, I have spent a lot of time talking to educators, researchers, and engineers about AI. The feeling is almost universal: it is impressive, it is helpful, but you should absolutely not trust it with your math even if it sounds confident.

That tension between how capable AI feels and how accurate it actually is has been on my mind for months. AI is not going away. The challenge now is figuring out how to make it reliable.

That is where Maple MCP comes in.

Maple MCP (Model Context Protocol) connects large language models like ChatGPT, Claude, Cohere, and Perplexity to Maple’s world-class math engine.

When your AI encounters math, your AI can turn to Maple to handle the computation so the results are ones you can actually trust.

It is a simple idea, but an important one: Maple does the math and the AI does the talking. Instead of guessing, the AI can be directed to call on Maple whenever accuracy matters.

Model Context Protocol (MCP) is an emerging open standard that allows AI systems to connect to external tools and data sources. It gives language models a structured way to request computations, pass inputs, and receive reliable outputs, rather than trying to predict everything in text form.

Here is a high-level view of how MCP fits into the broader ecosystem:

MCP lets an AI system connect securely to specialized services, like Maple, that provide capabilities the model does not have on its own.

If you want to learn more about the MCP standard, the documentation is a great starting point: Model Context Protocol documentation

Here is a glimpse of what happens when Maple joins the conversation:

Depending on the prompt, Maple MCP can evaluate expressions symbolically or numerically, execute Maple code, expand or factor expressions, integrate or solve equations, and even generate interactive visualizations. If you ask for an exploration or an activity, it can create a Maple Learn document with the parameters and sliders already in place.

As an example of how this plays out in practice, I asked Maple MCP:

“I'd like to create an interactive math activity in Maple that allows my students to explore the tangent of a line for the function f(x) = sin(x) + 0.5x for various values of x.”

It generated a complete Maple Learn activity that was ready to use and share. You can open the interactive version here: interactive tangent line activity .

In full disclosure, I did have to go back and forth a bit to get the exact results I wanted, mostly because my prompt wasn’t very specific, but the process was smooth, and I know it will only get better over time.

What is exciting is that this does not replace the LLM; it complements it. The model still explains, reasons, and interacts naturally. Maple simply steps in to do the math—the part AI cannot reliably do on its own.

We have opened the Maple MCP public beta, and I would love for you to try it.

Sign up today and we will send you everything you need to get started!

The inscribed square problem, also known as the Toeplitz conjecture, is an unsolved quastion in geometry: Does every plane simple closed curve (Jordan curve) contain all four vertices of some square? This is true if the curve is convex or piecewise smooth and in other special cases. The problem was proposed by Otto Toeplitz in 1911. For detailes see  https://en.wikipedia.org/wiki/Inscribed_square_problem

The Inscribed_Square procedure finds numerically one or more solutions for a curve defined by parametric equations of its boundary or by the equation F(x,y)=0. The required parameter of procedure  L  is the list of equations of the boundary links or the equation  F(x,y)=0 . Optional parameters:  N  and  R . By default  N='onesolution' (the procedure finds one solution), if  N  is any symbol (for example  N='s'), then more solutions.  R  is the range for the length of the side of the square (by defalt  R=0.1..100 ).

The second procedure  Pic  visualizes the results obtained.

The codes of the procedures:

restart;
Inscribed_Square:=proc(L::{list(list),`=`},N::symbol:='onesolution',R::range:=0.1..100)
local D, n, c, L1, L2, L3, f, L0, i, j, k, m, A, B, C, P, M, eq1, eq2, eq3, eq4, eq5, eq6, eq7, eq8, eq9, sol, Sol;
uses LinearAlgebra;
if L::list then
L0:=map(p->`if`(type(p,listlist),[[p[1,1]+t*(p[2]-p[1])[1],p[1,2]+t*(p[2]-p[1])[2]],t=0..1],p), L);
c:=0;
n:=nops(L);
for i from 1 to n do
for j from i to n do
for k from j to n do
for m from k to n do
A:=convert(subs(t=t1,L0[i,1]),Vector): 
B:=convert(subs(t=t2,L0[j,1]),Vector):
C:=convert(subs(t=t3,L0[k,1]),Vector): 
D:=convert(subs(t=t4,L0[m,1]),Vector):
M:=<0,-1;1,0>;
eq1:=eval(C[1])=eval((B+M.(B-A))[1]);
eq2:=eval(C[2])=eval((B+M.(B-A))[2]);
eq3:=eval(D[1])=eval((C+M.(C-B))[1]);
eq4:=eval(D[2])=eval((C+M.(C-B))[2]);
eq5:=eval(DotProduct(B-A,B-A, conjugate=false))=d^2;
sol:=fsolve([eq1,eq2,eq3,eq4,eq5],{t1=op([2,2,1],L0[i])..op([2,2,2],L0[i]),t2=op([2,2,1],L0[j])..op([2,2,2],L0[j]),t3=op([2,2,1],L0[k])..op([2,2,2],L0[k]),t4=op([2,2,1],L0[m])..op([2,2,2],L0[m]),d=R});
if type(sol,set(`=`)) then if N='onesolution' then return convert~(eval([A,B,C,D],sol),list) else c:=c+1; Sol[c]:=convert~(eval([A,B,C,D],sol),list) fi;
 fi; 
od: od: od: od:
Sol:=fnormal(convert(Sol,list),7);
print(Sol);
ListTools:-Categorize((X,Y)->`and`(seq(is(convert(X,set)[i]=convert(Y,set)[i]),i=1..4)) , Sol);
return map(t->t[1],[%]);
else
A,B,C,D:=<x1,y1>,<x2,y2>,<x3,y3>,<x4,y4>:
M:=<0,-1;1,0>:
eq1:=eval(C[1])=eval((B+M.(B-A))[1]):
eq2:=eval(C[2])=eval((B+M.(B-A))[2]):
eq3:=eval(D[1])=eval((C+M.(C-B))[1]):
eq4:=eval(D[2])=eval((C+M.(C-B))[2]):
eq5:=eval(LinearAlgebra:-DotProduct((B-A,B-A), conjugate=false))=d^2:
eq6:=eval(L,[x=x1,y=y1]):
eq7:=eval(L,[x=x2,y=y2]):
eq8:=eval(L,[x=x3,y=y3]):
eq9:=eval(L,[x=x4,y=y4]):
sol:=fsolve({eq1,eq2,eq3,eq4,eq5,eq6,eq7,eq8,eq9},{seq([x||i=-2..2,y||i=-2..2][],i=1..4),d=R});
eval([[x1,y1],[x2,y2],[x3,y3],[x4,y4]], sol):
fi;
end proc:

Pic:=proc(L,Sol,R::range:=-20..20)
local P1, P2, P3, T;
uses plots, plottools;
P1:=`if`(L::list,seq(`if`(type(s,listlist),line(s[],color=blue, thickness=2),plot([s[1][],s[2]],color=blue, thickness=2)),s=L), implicitplot(L, x=R,y=R, color=blue, thickness=2, gridrefine=3));
P2:=polygon(Sol,color=yellow,thickness=0);
P3:=curve([Sol[],Sol[1]],color=red,thickness=3):
T:=textplot([[Sol[1][],"A"],[Sol[2][],"B"],[Sol[3][],"C"],[Sol[4][],"D"]], font=[times,18], align=[left,above]);
display(P1,P2,P3,T, scaling=constrained, size=[800,500], axes=none);
end proc:

Examples of use:

The curve consists of a semicircle, a segment and a semi-ellipse (find 1 solution):

L:=[[[cos(t),sin(t)],t=0..Pi],[[t,0],t=-1..0],[[0.5+0.5*cos(t),0.8*sin(t)],t=Pi..2*Pi]]:
Sol:=Inscribed_Square(L);
Pic(L,Sol);

The procedure finds 6 solutions for a non-convex pentagon:

 L:=[[[0,0],[9,0]],[[9,0],[8,5]],[[8,5],[5,3]],[[5,3],[0,4]],[[0,4],[0,0]]]:
Sol:=Inscribed_Square(L,'s');
plots:-display(Matrix(3,2,[seq(Pic(L,Sol[i]),i=1..6)]),size=[300,200]);

For an implicitly defined curve, only one solution can be found:

L:=abs(x)+2*abs(y)-sin((2*x-y))-cos(x+y)^2=3:
Sol:=Inscribed_Square(L);
Pic(L,Sol);

               
See more examples in the attached file.

Inscribed_Square.mw