LiteRT
Project record
Maintained by Google (Google AI Edge)135
LiteRT is Google's on-device runtime for machine learning and generative AI models on Android, iOS, desktop, web, and IoT platforms, with CPU, GPU, and NPU acceleration. LiteRT is the new name for TensorFlow Lite; it runs models converted from PyTorch, TensorFlow, and JAX.156
- Website: Website (external site: developers.google.com)
- Repository: Repository (external site: github.com)
- Repository: LiteRT-LM repository (external site: github.com)
- License: LICENSE (external site: github.com)
- Release notes: Reviewed release (external site: github.com)
Availability and license
Overall availability
Documented as available to the general public. Access conditions and license terms may still apply.12
Availability is separate from permission: read the license before using or redistributing.
Component reuse rights
- 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
| Item | Status | Notes and evidence |
|---|---|---|
| Source codeIs the source code publicly readable? | Public | Public repository in the google-ai-edge GitHub organization.1 |
| DocumentationIs user documentation published? | Public | Official documentation on Google's developer site covers conversion, runtime APIs, hardware acceleration, and LiteRT-LM.5 |
| InstallationAre installation instructions or packages publicly available? | Public | The README points to platform integration guides and documents building Linux and Android artifacts from source with Docker, CMake, or Bazel. Google's renaming announcement says LiteRT packages are distributed through Maven, PyPI, and CocoaPods.16 |
| Supported platformsAre supported operating systems or hardware documented? | Public | The README's platform table lists CPU support on Android, iOS, Linux, macOS, Windows, web, and IoT; GPU support through OpenCL or OpenGL (Android), Metal (iOS and macOS), and WebGPU; and NPU support from several chipset vendors on Android and Linux, with Intel NPUs on Windows. Some accelerators are marked as coming soon.1 |
| Release statusAre versioned releases published? | Public | Reviewed release v2.2.0, published August 13, 2026. The README says LiteRT provides nightly builds and targets stable releases on a 6 to 8 week cadence.91 |
What it is useful for
Run and use notes
- LiteRT-LM's README documents a command-line tool (litert-lm run) that downloads LiteRT-LM model files from Hugging Face and runs them on Linux, macOS, Windows, or Raspberry Pi, and lists Python, Kotlin, and C++ APIs as stable.7
- The README says TensorFlow Lite packages and the tensorflow/lite directory are in maintenance mode, receiving only critical security and stability updates, and directs new code to LiteRT's Compiled Model API.1
Organization context
Provenance and derivatives
Google renamed TensorFlow Lite to LiteRT in September 2024; the README says LiteRT continues the legacy of TensorFlow Lite.61
- Derived from: TensorFlow Lite (part of TensorFlow) — LiteRT is the renamed continuation of TensorFlow Lite.
U.S. eligibility
Eligible · basis: U.S.-governed project
LiteRT is published in Google's google-ai-edge GitHub organization; its README calls it Google's on-device runtime, source files carry Google LLC copyright notices, and contributions require a Google contributor license agreement. Google is headquartered in the United States (see the Google record).134
Sources
This listing is not an endorsement, a safety assessment, or a federal approval.