Is Equana compatible with MATLAB code?

No. Equana has its own language (.eq files) and cannot run .m files. The syntax is Julia-inspired and feels familiar to MATLAB and Julia users, but existing MATLAB code must be rewritten. If you need drop-in .m compatibility, look at GNU Octave instead.

Is Equana really free?

Yes. Equana is currently in alpha and free to use. The tutorials run code in your browser without an account, and the full notebook needs only a free sign-in. There is no license fee and no server-side compute to pay for, because all execution happens on your own device.

Does my code or data get sent to a server?

No. All execution happens locally in your browser via WebAssembly. Equana makes zero server calls to run your code; notebooks can be stored locally or in the cloud, whichever you choose.

Equana vs MATLAB Online: a free, browser-based alternative for numerical computing

How Equana's free, on-device notebooks compare with MATLAB Online for matrices, linear algebra, and plotting.

If you're searching for a free MATLAB alternative that runs in the browser, Equana is worth a look — with one thing to know up front: Equana is not MATLAB-compatible. It's a different language, built for the same jobs — matrices, linear algebra, FFTs, statistics, and plotting — with syntax that will feel familiar if you've used MATLAB or Julia. Unlike MATLAB Online, Equana executes everything on your device via WebAssembly: no license, no install, and your data never leaves your machine. Equana is currently in alpha and free to use.

	Equana	MATLAB Online
Cost	Free (currently in alpha)	Requires a MATLAB license or subscription; limited free tier
Install	None — runs in any modern browser; optional desktop app	None for the online version; desktop MATLAB requires installation
Where code runs	Entirely on your device (WebAssembly)	On MathWorks cloud servers
Privacy	Data never leaves your machine	Code and data are uploaded and processed in the cloud
Language	Equana's own `.eq` language — Julia-inspired, 1-based indexing, multiple dispatch. Not MATLAB-compatible	MATLAB language; runs existing `.m` code unchanged
Ecosystem	20 packages, 282 functions (linear algebra, sparse matrices, FFT, statistics, plotting). Young and growing	Decades of development, vast toolbox catalog. Far deeper today
Debugger	Integrated: breakpoints, stepping, variable inspection	Full MATLAB debugger
Plotting	Line, scatter, histogram, 3D surfaces, heatmaps, VTK volume rendering	Extensive, mature plotting
Maturity	Alpha	Production, commercial

In short: if you need Simulink, specific toolboxes, or must run existing .m files, MATLAB is the right tool. If you want a free, private, zero-install environment for the use cases people reach for MATLAB for — matrices, numerical math, plots — Equana covers a lot of that ground today.

A different language, built for the same jobs

Equana is its own language, not a MATLAB clone. The core is Julia-inspired — 1-based indexing, multiple dispatch, dynamic typing with optional annotations — so MATLAB and Julia users tend to feel at home quickly:

# comments start with #
f = x -> x^2                 # arrow functions
y = 2x + 3x^2                # implicit multiplication
println(sqrt(16.0))          # function calls
r = 1:2:10                   # ranges

The syntax stays close to the math you're expressing, with a small, consistent set of rules.

Everything runs on your device

MATLAB Online runs your code on MathWorks servers: everything you compute is uploaded first. Equana takes the opposite approach. The interpreter, the BLAS/LAPACK-style numerical kernels, and the plotting engine all run in your browser tab via WebAssembly, backed by a single shared linear memory — there are zero server calls during execution. For proprietary datasets, unpublished research, or regulated environments, that difference matters: nothing to upload means nothing to leak.

And local doesn't mean slow. On a Ryzen 9950X, a 1000×1000 double-precision matrix multiply takes about 649 ms in plain JavaScript, about 57 ms with Equana's WASM SIMD kernel, and about 15 ms with the multithreaded WASM kernel; at 4000×4000 the multithreaded kernel reaches roughly 410 GFLOPS on 32 threads. Core math is built in — no toolbox to license:

function r = deg2rad(x)
  r = x * pi / 180
end
arr = [3, 1, 4, 1, 5, 9]

What you give up (for now)

A useful comparison names the gaps. Equana has no Simulink equivalent, no toolbox catalog comparable to MathWorks', and no ability to run existing MATLAB code — migrating means rewriting. The good news is that everyday numerical code translates directly; a MATLAB loop becomes:

for x in arr
  s = s + x
end
if x < lo
  r = lo
else
  r = x
end

Equana ships 282 functions across 20 packages — dense and sparse linear algebra, NDArray operations, FFT, polynomials, complex numbers, special functions, statistics, and plotting — which covers everyday numerical work, but decades of MATLAB toolboxes it is not. And Equana is in alpha: expect rough edges, and don't bet a production workflow on it yet.

Try it

The fastest way to judge a language is to run it. The Equana examples execute code right in your browser with no sign-up and no install — open one and press run. When you're ready for full notebooks with saved files, plots, and the debugger, sign in free. Everything still runs on your own machine.

Equana