LiteRT Studio is a powerful, completely offline AI chat application built for Linux. It provides a seamless, self-contained environment to run Large Language Models (LLMs) and Vision-Language Models locally on your machine without an internet connection.
- 100% Offline: No internet required once models are downloaded. Your privacy is guaranteed.
- Built-in Model Store: Browse, download, and manage LiteRT-optimized models directly within the application.
- Universal Hardware Support: Optimized for CPU inference, allowing it to run on almost any hardware (including older Laptops/Desktops).
- Zero Configuration: No need to worry about Python environments, CUDA, or complex dependencies. It just works out of the box.
- Vision Model Support: View and interact with images using vision-language models.
- Resource Control: Dedicated CPU thread allocation settings to balance performance and system usage.
- Modern UI: A premium, responsive interface featuring glassmorphism aesthetics and animated transitions.
Currently, LiteRT Studio is distributed as a .deb package for Debian-based distributions (Ubuntu, Linux Mint, etc.).
- Download the latest
.debpackage from the Releases page. - Install it using
dpkgor your favorite package manager:sudo dpkg -i litert_studio_1.0.0_amd64.deb
- Launch "LiteRT Studio" from your application menu.
LiteRT Studio uses the LiteRT-LM runtime to execute models efficiently on CPUs. It packages all necessary shared libraries and binaries, ensuring a standalone experience regardless of your system's global environment.
Note: This repository currently hosts the binary releases. Source code integration is planned for a future update.
- PS: Check new releases for fixes and feature updates