MLX

license

MIT(OSI)

source

public(ml-explore/mlx)

core features withheld

no — no enterprise or ee path in the repository root and no paid tier described in the README

MLX is released under the stock MIT license, with no appended condition, and the full source is public in this repository. The README describes the whole library with no enterprise or ee path and no paid tier, so the core is ungated.

• https://raw.githubusercontent.com/ml-explore/mlx/main/LICENSE recorded 2026-10-06

  LICENSE body, read in full: the stock MIT licence text at 1,066 bytes, carrying no appended term or additional condition.

• https://raw.githubusercontent.com/ml-explore/mlx/main/README.md recorded 2026-10-06

  README, read for the distribution and gating question: it documents installing from the mlx package or building from this repository, and names no enterprise, paid or licence-gated build.

Adoption is measured as PyPI downloads of the mlx package. GitHub stars are also recorded, but only as corroboration, since a star is not an install.

• https://api.github.com/repos/ml-explore/mlx recorded 2026-08-31

  Repo metadata for ml-explore/mlx: stargazers_count 28237, archived false, pushed_at 2026-08-30.

• https://pypistats.org/api/packages/mlx/recent recorded 2026-08-31

  last_month 1464066, last_week 250585, last_day 32853 for the mlx package.

MLX has the shape of a foundational framework - its own array type, composable transforms and autodiff, and an ecosystem built on it - but it runs on Apple silicon alone, where PyTorch, TensorFlow and JAX are hardware-portable. OpenVINO sets the precedent here: a technically strong framework scoped to one vendor's hardware.

• https://raw.githubusercontent.com/ml-explore/mlx/main/README.md recorded 2026-10-07

  The README's key-features list documents a NumPy-following Python API with C++, C and Swift APIs, higher-level mlx.nn and mlx.optimizers packages following PyTorch, composable function transformations for automatic differentiation, automatic vectorization and graph optimization, lazy computation, and dynamic graph construction.