Affiliate disclosure: ToolBistro may earn a commission from some links, at no extra cost to you. Facts come from official sources; we do not publish fabricated testing or ratings.

AI Tools

Is Woxi a Real Free Mathematica Alternative?

Woxi, a free open-source Mathematica alternative written in Rust, covers a working subset of the Wolfram Language and passed about 26,000 tests as of August 2026, while Mathematica commercial licenses start at $1,910 per year. Here is how Woxi compares with other Mathematica alternatives, verified on the official pages.

Key facts

What matters

  • Woxi is an AGPL-3.0 Rust interpreter for the Wolfram Language with a notebook editor (Woxi Studio), CLI, Jupyter kernel, Python package, and WASM build (GitHub, verified Aug 13 2026).
  • Mathematica commercial subscriptions cost $1,910 per year; the perpetual Plus tier is $3,820 with an estimated $999 per year renewal (official Wolfram pricing page, Aug 2026).
  • Personal Mathematica use costs $195 per year ($390 for the perpetual tier with a $175 renewal), so the price gap behind most mathematica alternatives searches is real.
  • Woxi conformance is checked by about 26,000 unit tests plus about 900 Wolfram script snapshot tests, all expected to match WolframScript output.
  • The honest catch: Woxi implements a subset of the language, ships no curated Wolfram data or AI Access, and AGPL-3.0 copyleft limits commercial embedding.

What is Woxi?

Woxi is an open-source interpreter for the Wolfram Language, the language behind Mathematica, reimplemented in Rust. The project has been public since September 2024 and sits at about 768 GitHub stars as of August 13, 2026, with the repository updated the day before this article was written.

The Show HN post that put Woxi on the Hacker News front page on August 12, 2026 lists what ships with it: Woxi Studio, a Mathematica-like notebook GUI built with the iced toolkit; a command-line interpreter; a Jupyter kernel; a Python package on PyPI; an npm package; and a WebAssembly module that runs in the browser. A JupyterLite demo runs fully in the browser with no cloud backend.

Woxi's defining trait is speed. The interpreter starts in milliseconds rather than the seconds a Wolfram kernel needs, because there is no kernel process to spawn and no license to verify. That makes it practical for shell scripts, one-liners, and short-lived processes.

How much does Mathematica cost in 2026?

Mathematica has no free tier. The official Wolfram pricing pages, verified on August 13, 2026, show a commercial Professional subscription at $1,910 per year, while the Professional Plus tier costs $3,820 and includes a perpetual desktop license with an estimated $999 per year renewal.

Personal use is cheaper but still priced: $195 per year for the standard tier, or $390 for Personal Plus with the perpetual desktop license and a $175 estimated renewal. Every tier includes the current Wolfram Language, two activation keys, cloud access, Wolfram AI Access Basic, and Wolfram Local MCP. An upgrade to Wolfram AI Access Pro adds $240 per year.

All online purchases are set to auto-renew, and each license permits installation on up to two computers. If the subscription lapses, only the Plus tiers keep desktop access working, and cloud access, upgrades, and premium support drop away with it.

Woxi vs Mathematica: what actually carries over

Woxi's own comparison page is candid: it documents missing features by Mathematica release, from version 1.0 (1988) through version 15.0 (2026). The project targets a usable subset for CLI scripting and notebooks, and its test suite of about 26,000 unit tests plus about 900 Wolfram script snapshot tests is run against WolframScript output, so what is implemented is verified rather than guessed.

Things Woxi does not carry over: full Wolfram Language coverage, the curated Wolfram Knowledgebase (financial, weather, geographic, and scientific datasets), Wolfram AI Access, cloud sync, and Mathematica's interactive Dynamic and Manipulate workflows. The desktop notebook experience is also a young open-source implementation, not a 36-year-old commercial product.

Licensing is the other real difference. Mathematica is proprietary and subscription-based. Woxi is AGPL-3.0, which is free to use but copyleft: any distributed modification must stay AGPL, a constraint that matters for commercial embedders.

Who should switch to Woxi (and who should not)

Switch if:

  • You script the Wolfram Language on a command line and want startup in milliseconds instead of seconds.
  • You are a student, researcher, or hobbyist priced out of the $195 per year personal license.
  • You want a Wolfram-Language-compatible engine embedded in a browser via WASM or inside another application.
  • You work from notebooks that use core language functions covered by the test suite.

Stay on Mathematica if:

  • Your work depends on curated data sources, Wolfram AI Access, or the long tail of built-in functions.
  • You need enterprise support, priority help, or a commercial support contract.
  • Your company would distribute a modified version and cannot accept AGPL-3.0 copyleft.
  • You rely on notebook portability with teams that all run Mathematica.

Other free Mathematica alternatives worth knowing

If Woxi's subset does not cover your needs, the other serious open-source option is Mathics, an open-source Mathematica kernel written in Python. The active repository is Mathics3/mathics-core, licensed GPL-3.0, with about 1,200 stars and commits as recent as August 10, 2026. The original mathics/Mathics repository is archived and points to the new home.

The practical difference between the two is platform: Woxi is a standalone Rust interpreter with its own notebook editor, CLI, and WASM build, while Mathics is a Python kernel whose core repository contains the Python modules implementing Wolfram Language built-in functions. Neither replaces Mathematica's curated knowledgebase or its full function set, so the decision is a trade between a $1,910 annual subscription and accepting a subset.

How to try Woxi today

Woxi installs with cargo install woxi from the Rust toolchain, and every GitHub release ships prebuilt archives for Linux, macOS, and Windows on x86-64 and arm64, with a SHA256SUMS.txt checksum file. The macOS build of Woxi Studio is not notarized, so macOS quarantines it after download until you clear the flag with xattr.

Python users can run pip install woxi, Node.js users can npm install woxi-wasm, and anyone with a browser can open the JupyterLite demo on woxi.ad-si.com without installing anything. The CLI accepts eval, run, and repl subcommands, and woxi run notebook.nb executes notebook files directly, which WolframScript itself cannot do.

At a glance

Woxi (open source)Mathematica (Wolfram)
PriceFree, AGPL-3.0From $195/yr personal; $1,910/yr commercial
Language coverageWorking subset, 26,000+ testsFull Wolfram Language
Startup timeMillisecondsSeconds (kernel plus license check)
InterfacesWoxi Studio, CLI, Jupyter, Python, npm/WASMDesktop + Cloud, WolframScript
Curated data and AINot includedKnowledgebase; AI Access Basic included, Pro +$240/yr
License typeCopyleft (AGPL-3.0)Proprietary subscription
Best forScripts, notebooks, embedding, budgetsResearch, data-heavy workflows, teams

FAQ

Is Woxi free to use?

Yes. Woxi is AGPL-3.0 open source, installed with cargo install woxi or from prebuilt GitHub release binaries, and it has no paid tier. Verified on the repository on August 13, 2026.

Can Woxi open Mathematica notebooks?

Partially. The CLI runs .nb files with woxi run notebook.nb, which WolframScript cannot do, and Jupyter notebooks work including graphical output. Full notebook compatibility is not complete because Woxi implements a subset of the Wolfram Language.

What are the best free Mathematica alternatives?

For Wolfram Language syntax, Woxi (Rust, AGPL-3.0) and Mathics (Python, GPL-3.0) are the two realistic open-source options. Woxi targets CLI scripting, notebooks, and embedding; Mathics is a Python kernel. Both implement a subset, so neither ships Mathematica's curated data, AI features, or full function coverage.

Related reading

Radar index

Sources