Holaos
Turn repeat work into running AI work-streams.
Last verified:
What is Holaos?
holaOS is an open-source Agent Computer that takes on roles for you, designed as a shared environment where humans and AI agents operate side by side. Unlike traditional AI agents that reset after each session, holaOS provides persistent memory, allowing agents to keep memory, use tools, run apps, and continue work across runs. It reimagines the computer as a collaborative space where AI teammates learn your goals, adapt to your routine, and keep getting better with you.
Key features include Multi Workspaces for isolating different workstreams with their own context, rules, tools, and history; Sub Agents that handle complex tasks in parallel while you interact with a centralized coordinator; Dashboard for tracking what's running and identifying what needs review; and the Kickoff flow that transforms recurring work into structured goals, context, rules, sources, and first artifacts. The system uses Model Context Protocol (MCP) for tooling and runs as an Electron-based desktop application.
holaOS is ideal for professionals dealing with recurring, context-heavy work-streams such as weekly research, content production, customer feedback management, launch planning, and client delivery. It targets developers, researchers, content creators, project managers, and anyone who needs AI teammates for long-horizon work that unfolds over time. The open-source MIT license allows users to inspect the runtime, build apps on top of it, or self-host according to their workflow needs.
Holaos pricing
Pricing model: Freemium
holaOS is completely free and open-source under MIT license. Users can download, install, and self-host without any cost. The project is available on GitHub with 5.7k+ stars. No paid tiers or subscriptions exist - all features including Multi Workspaces, Sub Agents, and Dashboard are included in the free open-source version. Users only need to provide their own LLM API keys or use their preferred AI provider.
Holaos pros
- Open-source under MIT license for full transparency
- Persistent memory that survives sessions and doesn't reset
- Shared environment with same browser, files, and apps as user
- Multi Workspaces for isolated context per workstream
- Sub Agents handle complex tasks in parallel
- Kickoff flow structures work into goals, rules, and artifacts
- Dashboard tracks running tasks and next steps
- Agents learn and improve from user corrections over time
- Self-hostable for privacy and workflow control
- Built-in runtime continuity across interruptions
- Portable and inspectable operating unit
- Supports Model Context Protocol (MCP) for tooling
- Electron-based desktop application with GUI
- Compounds value with each run through visible rules
- Designed specifically for long-horizon recurring work
Holaos cons
- Currently in Beta 0.1, still early stage
- macOS support primary, Windows and Linux in progress
- Requires Node.js 24.14.1 and npm installation
- Learning curve for workspace and kickoff setup
- Self-hosting requires technical expertise
- Experimental feature with potential bugs
- Limited to desktop environment, no mobile app
- Depends on user having LLM keys or providers
Frequently asked questions about Holaos
What is holaOS?
holaOS is an open-source Agent Computer that takes on roles for you. It is not another AI agent but a shared environment where you and your AI teammate operate side by side, with full access to the same browser, files, and apps. Agents keep memory, use tools, run apps, and continue work across runs without resetting.
How is holaOS different from regular AI agents?
Unlike regular AI agents that are stateless and forget everything when you close the chat, holaOS provides persistent memory that survives sessions. It creates living workspaces with goals, rules, tools, artifacts, history, and next runs. The workspace compounds value - corrections you make become visible rules so the next run starts smarter.
What is the Kickoff flow?
The Kickoff flow is what makes holaOS different from a blank agent builder. You give it a goal, context, sources, rules, and output format. It structures the work and produces the first output right away. When you review and correct it, those corrections become visible rules for future runs.
What are Multi Workspaces?
Multi Workspaces in Beta 0.1 give each long-running workstream its own isolated context, rules, tools, and history for better organization and focus. Each workspace is self-contained with its own agent identity, memory, installed apps, and standing instructions, isolated from every other workspace.
What are Sub Agents?
Sub Agents help complex work move forward in parallel - research, drafting, verification, synthesis - without manually juggling chats. You interact with a single centralized coordinator agent while Sub Agents handle different aspects of the task simultaneously.
How do I install holaOS?
Run the holaOS install script from the GitHub repository. It installs git and Node.js 24.14.1/npm if missing, clones the repo into ~/holaboss-ai, runs desktop:install, creates the .env file from the example, runs desktop:prepare-runtime:local and desktop:typecheck, and only runs desktop:dev if you ask for --launch. The basic command is npm run desktop:install.
Is holaOS free?
Yes, holaOS is completely free and open-source under the MIT license. You can dig into the runtime, build apps on top of it, or self-host if that fits your workflow. There are no paid tiers or subscriptions - all features are available for free.
What work is holaOS best for?
holaOS is best for recurring, context-heavy work-streams like weekly research, content production, customer feedback, launch planning, and client delivery. It excels at work that unfolds over time: research that changes every week, content that needs your voice, feedback that should become product decisions, launches with many moving parts, and client work needing isolated context.
What platforms does holaOS support?
holaOS is implemented in TypeScript and runs as an Electron desktop application. It currently supports macOS with Windows and Linux in progress. The project is actively developing cross-platform support.
Can I self-host holaOS?
Yes, holaOS is open-source under MIT, so you can self-host it if that fits your workflow. The canonical configuration path is apps/desktop/.env where you can set values for your provider and machine. Self-hosting gives you full control over your AI environment and data privacy.