selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

MachinaOS logo

Self-host MachinaOS

v0.3.1

AI organisation with six self-improving employees

BetaOpen SourceMIT
Updated8 hours ago
LanguagePython
Setup difficultyAdvanced
ThemeDark mode
Stars
1 042
CPU
2 CPU cores
RAM
2GB RAM
Storage
10GB disk space
Port
5678 (Main app (API + WebSocket + SPA))
Install with Docker Compose

What is MachinaOS?

A self-hosted AI platform that runs six specialised digital workers covering software development, visual design, financial accounting, sales outreach, marketing campaigns, and tax preparation — each taking full ownership of its domain. You describe tasks in natural language and the system assembles a lead agent with specialists, connecting them to over 140 tools like email, calendars, code execution, messaging, payments, a browser agent, and Android device control. Work runs on a configurable token budget with support for 12+ model providers and local models. A visual canvas with multiple themes lets you see and edit workflows. Workers retain long-term memory and improve through editable skill playbooks. Install via desktop apps, a one-line script, Docker Compose, or from source. All data remains on your hardware with encrypted credentials; hosted tiers are optional.

Should you self-host MachinaOS?

MachinaOS scores 54/100 on selfhost.directory (how scoring works) and ranks #74 of 209 in AI Agents. It has 1 042 GitHub stars.

It shipped 14 stable releases in the past year. The latest, v0.3.1, came out on 23 October 2026. New versions land roughly every 1 days.

Setup is advanced: plan for several services and some ongoing maintenance. It is MIT-licensed and written mostly in Python.

Its closest alternative on selfhost.directory is AstrBot (88/100, 44 330 stars). It scores higher overall, so compare the two before you pick. See MachinaOS vs AstrBot side by side.

Features

  • 12+ LLM providers
  • 140+ tool integrations
  • 1Password credential bindings
  • AI employees with persistent memory
  • Android device controlWindows x64 only; iOS not supported
  • Browser AI agent
  • Durable execution with Temporal
  • Lead review of agent work
  • Plugin Architecture
  • Self-improving skills
  • Speech and translation
  • Token budgeting and cost tracking
  • Visual Workflow Builder
  • WhatsApp/Telegram/Discord/X automation

MachinaOS README

How to install MachinaOS

Install MachinaOS with Docker Compose

Docker and Docker Compose
Environment variables
OPENCOMPANY_OWNER_EMAIL, OPENCOMPANY_OWNER_PASSWORD, EXTRAS (browser, git)

git clone https://github.com/zeenie-ai/OpenCompany.git cd OpenCompany docker compose up -d --build

Open http://localhost:5678 and register the owner account. Optional: set OPENCOMPANY_OWNER_EMAIL and OPENCOMPANY_OWNER_PASSWORD in docker-compose.yml to create owner on first start.

Single container, all state in named volume; port published on 127.0.0.1 only by default

MachinaOS releases & version history

FAQ

What are the default ports?

Default ports (one continuous block): app 5678 (always what you open in the browser), dev-mode backend 5679, Temporal UI 5680 + gRPC 5681, Node.js executor 5682, WhatsApp 5683, plus Redis 6379.

How do I update OpenCompany?

Run git pull, then company clean and company start. The start command auto-installs any new dependencies.

Can I run OpenCompany on Windows?

Yes, OpenCompany runs natively on Windows with Node.js 22+ and Python 3.12+ installed. macOS, Linux, WSL2, and Git Bash are also supported.

What are the system requirements?

Minimum: 2GB RAM, 2 CPU cores, 10GB disk space. Recommended: 4GB+ RAM, 4+ CPU cores, SSD storage.

What does OpenCompany cost to run?

Free is $0 — you pay your model provider for tokens and your machine costs. Spend is metered against a budget you set. Pro and Max are hosted plans billed monthly on usage with no per-seat charge; price agreed on a call from the pricing page.

Where is my data stored?

On the Free plan, on your own infrastructure with no vendor-side copy. On hosted Pro and Max plans, data lives in your hosted workspace; Max runs as a dedicated isolated instance in your chosen region.

How is OpenCompany different from OpenClaw, Hermes Agent, or Paperclip?

Those are agent runtimes and coordination layers. OpenClaw is a personal assistant, Hermes Agent is a long-running agent runtime, and Paperclip is a governance layer. OpenCompany ships the organisation itself: fixed business roles, token budget in, revenue tracking out.

How is OpenCompany different from Zapier, Make, or n8n?

Those are workflow automation tools where you design every trigger and action. OpenCompany assigns an outcome to a role instead — the AI employee decides how to complete it and reports back.

Do I need to be a developer to use OpenCompany?

You need to be comfortable deploying self-hosted software. Once running, work is assigned in plain language — you do not build workflows step by step because each AI employee already owns its function.

Can I self-host OpenCompany?

Yes, and it is free to self-host. The Free plan runs on your own servers so documents, customer data, and financial records never leave infrastructure you control. Hosted Pro and Max plans are available if you prefer not to self-host.

Is OpenCompany free and open source?

Yes, the source is public on GitHub under the MIT licence. The Free plan is the whole product run on your own machine; you only pay for LLM tokens and infrastructure. Hosted Pro and Max plans are billed monthly on usage.

How does OpenCompany work?

You set a token budget. Work is dispatched to the AI employee that owns that function, which completes it and returns finished output. Each run reports token cost and revenue produced.

Who makes OpenCompany?

OpenCompany is built and maintained by Zeenie AI. The source is public on GitHub under the zeenie-ai organisation.

What is OpenCompany?

OpenCompany is an open-source, self-hosted AI system that runs the operational functions of a small business with six role-specialised AI employees — developer, designer, accountant, sales, marketer, and tax preparer — each owning a function end-to-end.

More in Business Software

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…