AI Potluck
Model components / Inference code

Uzu

Mirai Tech

Uzu is a high-performance on-device inference engine for Apple platforms, running models locally on iOS and macOS against unified memory so that inference needs no network round trip and no per-token cost. It ships Python, TypeScript and Swift bindings in-repo.

The Rust and Swift distribution is primary, so the PyPI wrapper is not declared as an artifact. Verified 2026-08-31 via GitHub, the LICENSE body and the repository README.

Openness

5 medium confidence
5.0
license
MIT(OSI)
source
public(the published repository is the engine)
core-gated
ungated(no enterprise path in the repository root and no paid build of the engine in the README)

The LICENSE body is the stock MIT text, Copyright (c) 2025 Mirai Tech Inc., with no appended condition, read in full rather than taken from the API's label. The repository is public and unarchived and builds the engine itself. Its root tree carries no enterprise, ee or commercial directory and its README describes no licence-gated build, so the core reads as ungated. Confidence is medium because that is a repository-and-README read rather than a pricing-page read.

Adoption

2 low confidence
2.0

1,684 GitHub stars, in the 1K-10K band of the stars scale. A `uzu` package exists on PyPI and does resolve to this repository, but it drew 788 downloads in the trailing 30 days against a Rust and Swift primary distribution, so banding on it would measure a minority channel rather than the product. No package artifact is declared for that reason.

Capability

3 medium confidence
3.0

Two bands below the vllm anchor, and level with llamafile, which is the other single-target local engine here. It serves one hardware family and does not carry the multi-node or parallelism surface the bands above are placed on.

Verified 2026-08-31