What You Need to Install and Run Kompas on macOS or Windows

Updated

To run Kompas on your own machine, download the desktop app for your platform — macOS (Apple Silicon or Intel) or Windows (x64) — and choose one of two setups. The Free desktop version is local-first and requires you to supply your own AI access (an OpenAI or Anthropic API key, a Claude Code or Codex subscription, or local models via Ollama). The Pro Cloud version skips that setup with built-in AI and cross-device sync, and starts with a 7-day free trial that the site says requires no credit card.

Download the Right Build

Kompas publishes separate installers for each platform and chip type. Pick the one that matches your machine:

Your machine Download option
Mac with Apple Silicon (M-series) Download for macOS · Apple Silicon
Mac with Intel processor Download for macOS · Intel
Windows PC (64-bit) Download for Windows · x64

The site lists the free desktop app as available for both macOS and Windows and as open source under MIT. If you are unsure which Mac build you need, check your chip under → About This Mac before downloading.

Choose Your Setup Path

Both paths use the same desktop app; the difference is who handles the AI and where your ideas live.

Free desktop (local-first, bring your own AI)

This is the option to pick if you want your ideas to stay on your device and you already have — or don't mind setting up — an AI provider. According to the site, the free desktop version lets you:

  • Keep your ideas on your own device (local-first).
  • Connect your own OpenAI or Anthropic key.
  • Use a Claude Code or Codex subscription you already pay for.
  • Run local models through Ollama.

The trade-off is setup: you configure the AI connection yourself, and there is no cross-device sync described for this tier. The app is open source under MIT, so the desktop software itself is free to use.

Pro Cloud (built-in AI, no setup)

Choose Pro if you would rather not manage API keys or model configuration. The site describes Pro as including:

  • Cross-device sync between your desktops.
  • Frontier AI built in — no setup, no API keys.
  • Included AI usage with clear limits and optional top-ups.
  • A 7-day free trial with no credit card required.

The site's own guidance is to "start free" and move to Pro when you want sync and AI handled for you.

Step-by-Step: Get Running

  1. Confirm your platform. Note whether you are on Apple Silicon, Intel Mac, or Windows x64.
  2. Download the matching installer from the options above.
  3. Install and open the app. On macOS, you may need to allow the app in System Settings → Privacy & Security the first time you launch a downloaded app.
  4. Pick your AI path.
    • Free: open settings and add your OpenAI/Anthropic key, sign in with a Claude Code or Codex subscription, or point Kompas at a local Ollama model.
    • Pro: start the 7-day trial and use the built-in AI without adding keys.
  5. Verify it works. Start a session and send a prompt. If the AI responds, your connection is live. If not, the issue is almost always the AI step — see below.

Common Stumbling Blocks

  • No AI response on the free version. The desktop app does not ship with a model. Without a key, subscription, or local model, capture may work but AI-assisted sessions will not. Recheck your provider settings.
  • Wrong Mac build. An Apple Silicon installer will not run properly on an Intel Mac, and vice versa. Re-download the correct one.
  • Ollama not detected. Local models require Ollama to be installed and running on the same machine before Kompas can use it.
  • Expecting sync on Free. Cross-device sync is described as a Pro feature; on the free desktop version, your ideas stay on that device.

What You Get Once It's Running

Kompas is built around sessions that become a connected graph, which then feeds your writing. The site describes the flow as capture → connect → write: your reading, conversations, and half-formed ideas are gathered in one place, linked to earlier notes, and turned into drafts. The app's workspace includes Capture, Sessions, Projects, Library, Writer, and Graph areas, so once installation and AI setup are done, the first useful step is to run one session and let it land in your graph rather than starting from a blank page.