local-ai

LiteRT: Fast Edge AI

Name: LiteRT: Fast Edge AI
Author: google-ai-edge

Deploy high-performance ML models on edge platforms with LiteRT, for startup founders, with 2.5k+ GitHub stars

From the repo · “LiteRT, successor to TensorFlow Lite. is Google's On-device framework for high-performance ML & GenAI deployment on edge platforms, via efficient conversion, runtime, and optimization”

advanced⏱ Half a day💵 Free (self-hosted)

3,518 stars+22★ this week478 forksC++Guide quality 8/10 · reviewed Jun 29Updated 10/22/2026100% free · open source

What it is

Run ML and GenAI models directly on edge devices with high performance.

Real use case example

"A founder building a smart home device wants to recognize objects and provide information to users. They write a LiteRT model to interpret camera feeds and integrate it with their system; next, they test and refine it to achieve high accuracy; finally, they deploy the working model on their edge device to start improving user experiences."

Best insideClaude DesktopClaude CodeAny AI Client

When to use it

•When you need to run ML models on resource-constrained edge devices
•When you want to optimize the performance of your ML models on-device
•When you need to deploy ML models on platforms with limited connectivity

Ready-to-paste prompt

bazel build -c opt //examples:image_classification:main --define lite=1 && bazel run -c opt //examples:image_classification:main --define lite=1 -- --model_file=model.lite

Heads up: Ensure you have Bazel installed and properly configured on your system, as LiteRT relies on Bazel for building and deploying models

Saves to your device

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/litert && curl -fsSL https://workflowstacks.com/api/skills/litert/claude-skill -o ~/.claude/skills/litert/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How LiteRT: Fast Edge AI works

Codeflow

Free to inspect

LiteRT: Fast Edge AI is a very large C++ project (~1.3M lines across 5386 code files, plus 356 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, Apache-2.0 license, has a test suite.

Size

Very large codebase

~1.3M lines · 5386 code files · days to read — use, don't read

Setup

Developer setup

A real software project. Use it via its install path; don't expect to read it all.

Runs on

C++

No API keys detected

C++ 82%Python 5%Starlark 5%C 2%Java 2%

Where to start reading

1
README.md
Start here — what it does and how to install it
2
litert/__init__.py
Inside litert/ — the main logic begins here

What's in each folder

litert/Core code — the actual logic2658 files

docs/Documentation14 files

tflite/Folder3406 files

ml_drift_delegate/Folder561 files

tensor/Folder178 files

ci/Folder44 files

docker_build/Folder10 files

cmake_example/Folder3 files

READMEHas testsDocumentedCI checksApache-2.0 licenseUpdated this month

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/litert && curl -fsSL https://workflowstacks.com/api/skills/litert/claude-skill -o ~/.claude/skills/litert/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

Quick Actions

Details

Creator

google-ai-edge

Language

C++

Related skills

More local-ai tools founders pair with this one.

local-ai★ 132,005

Run Powerful LLMs Locally with llama.cpp – Fast C++ Inference

Empower your product with on‑device LLM inference using llama.cpp’s efficient C++ engine. Ideal for founders who need privacy, speed, and no cloud costs. 130k+ GitHub stars.

local-ai★ 94,719

Stirling-PDF: Edit PDFs Anywhere

Edit PDFs on any device with Stirling-PDF, a tool for startup founders, backed by 80k+ GitHub stars

local-ai★ 38,207

Run AI Locally with llmfit

Find compatible AI models for your hardware with llmfit, a tool for founders, with 29k+ GitHub stars.

local-ai★ 8,452

GenieX: Run AI locally

Get local AI on Qualcomm devices with GenieX, a tool for startup founders, backed by 8.1k+ GitHub stars.

local-ai★ 5,866

Maka‑Agent: Boost Productivity with a Local‑First AI Desktop Assistant

Maka‑Agent brings AI-powered assistance directly to your desktop, running offline for privacy and speed—ideal for founders who need instant, secure help without cloud dependencies.

local-ai★ 3,532

OpenBiliClaw: Private AI Agent that Curates Multi‑Platform Content

OpenBiliClaw lets founders automatically surface relevant posts from Bilibili, Xiaohongshu, Douyin, YouTube and more, while keeping data fully local and private. Ideal for teams building recommendation features – 3.5k+ GitHub stars.