taOS: Self-Hosted AI Agent OS for Your Hardware and Data Sovereignty

Introduction

taOS, short for Tiny Agent OS, is a groundbreaking self-hosted AI agent operating system that puts you in control of your data and AI. It allows your AI agents' memory, conversations, and files to reside entirely on hardware you own, ensuring privacy and data sovereignty. Designed to run offline by default with cloud options available by choice, taOS transforms your existing consumer hardware, from Raspberry Pis to gaming PCs, into a powerful, distributed AI compute cluster.

Installation

Getting started with taOS involves a straightforward installation process for both the controller and worker nodes.

To install the controller (server) on Debian, Ubuntu, Fedora, Arch, Alpine, or macOS, use this one-line command:

curl -fsSL https://raw.githubusercontent.com/jaylfc/taOS/master/scripts/install-server.sh | sudo bash

This script is idempotent and safe to re-run. After installation, access the web desktop at http://your-host:6969 (or http://taos.local:6969 with mDNS).

To add a worker to your distributed compute cluster, use the following command, replacing http://your-server:6969 with your controller's address:

curl -fsSL https://raw.githubusercontent.com/jaylfc/taOS/master/scripts/install-worker.sh | sudo bash -s -- http://your-server:6969

Workers will print a pairing code that needs to be approved in the taOS Cluster app for security.

Examples

taOS provides a rich, integrated experience with a full web desktop and a wide array of features:

Web Desktop Environment: Access a complete browser-based desktop with a window manager, dock, launchpad, and 43 bundled apps, including messaging, mail, projects, and creative studios. It adapts seamlessly to mobile and tablet devices, offering a widget-first home screen and PWA support.
Multi-Framework Agent Chat: Experience framework-agnostic messaging where agents from different frameworks (e.g., OpenClaw, Hermes, SmolAgents, Langroid) can converse in shared channels, maintaining their history and connections regardless of the underlying framework.
Creative Studios: Unleash your creativity with dedicated studio apps like Images Studio for AI-powered image generation and editing, Game Studio for AI-assisted game creation, and LoRA Studio for archiving and managing LoRA models on your cluster.
Distributed Compute Cluster: Combine various devices, from old laptops and Raspberry Pis to gaming PCs and Android phones, into a unified AI compute mesh. taOS intelligently routes tasks to the best available hardware, maximizing your local AI capabilities.
Comprehensive AI Memory: The taOSmd memory system offers a robust, self-hosted knowledge graph for your AI's long-term memory, ensuring high accuracy in retrieval and generation without relying on external cloud databases.

Why Use taOS?

Choosing taOS means embracing a future where your AI is truly yours. Its core philosophy of "Your AI, your hardware, your data" translates into several compelling advantages:

Data Sovereignty: All your AI's memory, conversations, and files remain on your physical hardware, giving you complete control and privacy.
Local-First Approach: taOS is designed to run fully offline, leveraging your local compute resources and model catalogs. Cloud services are optional, not mandatory.
Hardware Flexibility: It supports a vast range of consumer hardware, from low-power SBCs like Raspberry Pi and Orange Pi to powerful gaming PCs and Apple Silicon Macs, allowing you to utilize devices you already own.
Rich Feature Set: Beyond basic AI agents, taOS offers a full web desktop, an app store, creative studios, a sophisticated memory system, and robust monitoring tools, creating a complete ecosystem for personal AI.
Open Source and Auditable: Licensed under AGPL-3.0, taOS is open source, providing transparency and the freedom to inspect, modify, and self-host without vendor lock-in.