---
id: ai-edge-litert-nightly
version: "2.2.0.dev20260814"
license: Apache 2.0
license_treatment: permissive
maintenance: active
---

# ai-edge-litert-nightly — LiteRT is for mobile and embedded devices.

License: permissive · Maintenance: active · Downloads: 184.7K/mo

## What it is and what it does

LiteRT is Google's official runtime for executing machine learning models on mobile and embedded devices. It provides optimized inference with low latency and a small footprint, enabling on-device ML without cloud dependencies. The package wraps a compiled runtime and depends on numpy, protobuf, flatbuffers, and related utilities for tensor manipulation and model loading.

This is a nightly development build, meaning it receives daily updates and may include experimental features or breaking changes. It targets developers building ML applications for Android, iOS, and other edge platforms who need to run inference directly on the device rather than through a server.

Use it for:

- Run trained TensorFlow Lite models for inference on mobile phones and tablets without network latency.
- Deploy ML models to embedded systems and IoT devices with constrained memory and CPU.
- Build real-time computer vision or audio processing applications that execute locally on-device.
- Test and validate model behavior in a nightly development environment before stable release.
- Integrate pre-trained models into Android or iOS applications using Python bindings.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

LiteRT is a runtime for running trained machine learning models on mobile and embedded devices with low latency and minimal binary footprint.

Yes, if you are actively developing or testing ML inference on mobile/embedded platforms and are comfortable with nightly builds. The package is actively maintained, has no known vulnerabilities, and offers permissive licensing. However, avoid it for production deployments requiring stability—use a stable release instead. The medium install friction is acceptable given the precompiled wheels for common platforms.

## Install

pip install ai-edge-litert-nightly
uv add ai-edge-litert-nightly
poetry add ai-edge-litert-nightly

## Installing ai-edge-litert-nightly

Before you install: This is a nightly development build released daily. Medium install friction reflects precompiled wheels for Python 3.10–3.14 on Linux and Windows, but the dev version status means it may contain breaking changes or incomplete features compared to stable releases.

License in practice: Licensed under Apache 2.0, a permissive open-source license that allows commercial use, modification, and distribution with minimal restrictions.

Quickstart:

pip install ai-edge-litert-nightly

import litert

# Load and run a model on device
model = litert.Interpreter(model_path='model.tflite')
model.allocate_tensors()
model.invoke()

Requires a compatible Python version (3.10–3.14 based on available wheels) and a supported operating system (Linux manylinux_2_27 or Windows x86_64).

Verify before relying:

- Whether this nightly build is intended for production use or development/testing only
- Stability guarantees and breaking-change policy for nightly releases
- Performance characteristics and supported model architectures
- Exact Python version support (requires_python is unspecified)

## Package facts

- License: Apache 2.0 (permissive)
- Python support: unspecified
- Install friction: medium
- Maintenance: active
- Downloads: 184.7K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

mobile machine learning inference, on-device ml model execution, embedded neural network runtime, tensorflow lite runtime, low-latency edge inference, lightweight model deployment, android ios ml inference, edge-inference, mobile-ml, nightly-build

[View on SkillFed](https://skillfed.io/packages/ai-edge-litert-nightly) · [View on PyPI](https://pypi.org/project/ai-edge-litert-nightly/)