SoftwareRuntime

ExecuTorch

Project record

Maintained by Meta2 3

ExecuTorch is PyTorch's open-source stack for running AI models on-device, from phones and laptops to embedded systems and microcontrollers. A model is captured with torch.export, lowered for selected hardware backends into a .pte program, and run through C++, Python, Swift/Objective-C, Kotlin/Java, or JavaScript APIs.1 4Fact reviewed Oct 1, 2026

Website: Website (external site: executorch.ai)
Repository: Repository (external site: github.com)
Documentation: Documentation (external site: docs.pytorch.org)
License: LICENSE (external site: github.com)
Release notes: Reviewed release (external site: github.com)

Last reviewedOct 1, 2026Entry updated Oct 1, 2026Documented release Unknown

Availability and license

Overall availability

Public

Documented as available to the general public. Access conditions and license terms may still apply.1 2

Availability fact review: Oct 1, 2026

Availability is separate from permission: read the license before using or redistributing.

BSD-style license (3-clause) (external site: github.com)2 1
Applies to: Code · SPDX BSD-3-Clause · on the rubric’s OSI-approved software-license list · fact reviewed Oct 1, 2026

The LICENSE is a three-clause BSD license naming Meta Platforms, Inc. and affiliates plus Arm, Qualcomm Innovation Center, Apple, MediaTek, NXP, Samsung Electronics, and Intel as copyright holders. The README notes that embedded toolchains, simulators, and target runtimes are installed separately; this entry does not cover them.2 1Fact reviewed Oct 1, 2026

Component reuse rights

A readable or downloadable component is not automatically reusable. These indicators concern recorded license evidence, not system certification.

weights: Unknown — no complete fact-level rights review
code: Reviewed qualifying license recorded — check scope and conditions
data: Unknown — no complete fact-level rights review
documentation: Unknown — no complete fact-level rights review

No complete system-rights review is recorded for this release.

Public materials checklist

Items for a runtime under USASI rubric v0.2. Unknown means unassessed or insufficient evidence.

Public materials checklist for ExecuTorch
Item	Status	Notes and evidence
Source codeIs the source code publicly readable?	Public	Public repository in the pytorch GitHub organization.1
DocumentationIs user documentation published?	Public	Official documentation covers getting started, edge platforms, hardware backends, and LLM export and deployment.4
InstallationAre installation instructions or packages publicly available?	Public	`pip install executorch` in Python 3.10 to 3.14, with wheels for Linux x86-64, Linux AArch64, macOS arm64, and Windows x86-64; an Android AAR on Maven Central, Apple frameworks through Swift Package Manager, and source builds are also documented.1
Supported platformsAre supported operating systems or hardware documented?	Public	The README maps targets to backends: Android, iOS and iPadOS, macOS, Linux, Windows, browser/WebAssembly, and embedded microcontrollers, with backends including XNNPACK, Vulkan, Core ML, Qualcomm, MediaTek, Samsung Exynos, OpenVINO, Arm Ethos-U, NXP eIQ Neutron, and Cadence DSP. Several desktop and browser paths are marked experimental.1
Release statusAre versioned releases published?	Public	Reviewed release v1.5.1, published September 23, 2026. The README says it tracks main and points users to the latest release with the stable documentation.6 1

What it is useful for

Deploying PyTorch models, including quantized language, speech, and vision models, to run locally on mobile, desktop, browser, and embedded targets using CPU, GPU, NPU, or DSP backends.1Fact reviewed Oct 1, 2026

Run and use notes

Documented facts only. No hardware or performance claims are made without a cited source and stated assumptions.

The README lists local LLM building blocks, including quantized text models, tool calling, speculative decoding, and an experimental OpenAI-compatible local server example.1
A .pte program exported for one hardware backend is not a universal model file; targeting Core ML, Qualcomm, or another accelerator requires that backend's dependencies and a separate export, per the README.1

Organization context

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S.-governed project

ExecuTorch is developed in the pytorch GitHub organization. Its LICENSE lists Meta Platforms, Inc. first among the copyright holders, and its contributing guide requires the contributor license agreement used for Meta's open-source projects. The PyTorch Foundation website lists ExecuTorch among its projects. Meta and the PyTorch Foundation's host, The Linux Foundation, are U.S.-based (see their records). Other companies named in the LICENSE contributed hardware-backend code; this assessment rests on Meta's documented role.2 3 1 5

Assessed Oct 1, 2026

Sources

1.
pytorch/executorch README (external site: raw.githubusercontent.com)
PyTorch (Meta) · Repository · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026
2.
pytorch/executorch LICENSE (external site: raw.githubusercontent.com)
PyTorch (Meta) · License · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026
3.
pytorch/executorch CONTRIBUTING.md (external site: raw.githubusercontent.com)
PyTorch (Meta) · Repository · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026
4.
ExecuTorch documentation (stable) (external site: docs.pytorch.org)
PyTorch · Documentation · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026
5.
ExecuTorch (PyTorch Foundation project page) (external site: pytorch.org)
PyTorch Foundation · Official page · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026
6.
Release v1.5.1 (external site: github.com)
PyTorch (Meta) · Release notes · published Sep 23, 2026 · accessed Oct 1, 2026 · evidence reviewed Oct 1, 2026

Email a correction Report a correction Edit this entry

This listing is not an endorsement, a safety assessment, or a federal approval.