Numba Runs in the Browser: 250x Speedup in JupyterLite
Numba in the Browser: Unlocking a New Scientific Python Stack in JupyterLite
JupyterLite brings a full scientific Python stack to the browser via WebAssembly, but Numba was missing. Now, QuantStack has made it work: Numba's JIT compiler pipeline—from Python bytecode to LLVM IR to WebAssembly—runs entirely client-side, delivering a 249x speedup over standard Python. This unlocks PyTensor, PyMC, and other Numba-dependent libraries in JupyterLite, with packages available via emscripten-forge.
The larger relative gain makes Numba especially compelling in the browser, where bypassing Python interpreter overhead can have an even greater impact.