openTARS / README.md
MastLLM's picture
openTARS 5.3.5: cartão
3dddba9 verified
|
Raw History Blame Contribute Delete
2.72 kB
---
license: mit
tags:
- assistant
- local-ai
- desktop
- tool-use
language:
- en
- pt
---
# openTARS
<img src="https://huggingface.co/spaces/Mast-LLM/README/resolve/main/banner.png" alt="MastLLM, makers of openTARS" width="100%">
A local AI assistant for Linux, Windows and macOS. Ask in your own words and it opens programs, clicks buttons by name, types, searches the web, looks at the screen and runs commands. With a local model, processing stays on your machine. No subscription. Made by [MastLLM](https://huggingface.co/Mast-LLM).
This repository is a **mirror** of the app for people who find openTARS here. The source of truth, issues and releases are on GitHub: https://github.com/MastLLM/OpenTARS. Website: https://mastllm.github.io/OpenTARS/
## What is here
- `opentars_all.deb`: the Ubuntu/Debian package (version 5.3.5 "Tesseract", revision 2026112202).
- `app/`: the application source and the installers (`app/empacotamento/`).
- `LICENSE`: MIT.
## Install
Windows 10 or 11, in PowerShell:
```powershell
irm https://raw.githubusercontent.com/MastLLM/OpenTARS/main/app/empacotamento/windows/instalar.ps1 | iex
```
Or download [opentars-setup.exe](https://github.com/MastLLM/OpenTARS/releases/latest/download/opentars-setup.exe).
macOS 12 or newer, in Terminal:
```bash
curl -fsSL https://raw.githubusercontent.com/MastLLM/OpenTARS/main/app/empacotamento/macos/instalar.sh | bash
```
Or download the app: [Apple Silicon](https://github.com/MastLLM/OpenTARS/releases/latest/download/openTARS-macos-arm64.dmg) · [Intel](https://github.com/MastLLM/OpenTARS/releases/latest/download/openTARS-macos-x86_64.dmg).
Ubuntu, Debian and derivatives:
```bash
wget -O /tmp/opentars.deb https://github.com/MastLLM/OpenTARS/releases/latest/download/opentars_all.deb && sudo apt install -y --reinstall /tmp/opentars.deb
```
The app downloads its engine (llama.cpp) and Gargantua 2.5 the first time it opens, each checked against a fixed SHA-256. The install commands above download and run code from GitHub; read the script first if you like. How to verify the package: https://mastllm.github.io/OpenTARS/verify.html
## Models
- [Gargantua 2.5](https://huggingface.co/Mast-LLM/Gargantua-2.5): the main model (Gemma 4 E4B, with vision and audio).
- [Gargantua Lite Gen 3](https://huggingface.co/Mast-LLM/Gargantua-2.5-Lite): the smaller model (Gemma 4 E2B).
- [Gargantua Pro Gen 1](https://huggingface.co/Mast-LLM/Gargantua-Pro): the largest model, for missions in a virtual machine (Gemma 4 12B).
- [Gargantua Code Gen 2](https://huggingface.co/Mast-LLM/Gargantua-Code-Gen2): for programming.
- [Murph](https://huggingface.co/Mast-LLM/Murph): the tiny request classifier built into the app.