How Kompas Handles Privacy and Your Data
Updated
Kompas keeps your notes on your own device in its free desktop mode, and the company states that your data is never sold and never used to train models. If you want cross-device sync and built-in AI without managing keys, the Pro cloud plan moves your sessions off your machine, so syncing becomes the main privacy trade-off to weigh.
Free desktop: local-first by default
The free version of Kompas is a desktop app for macOS (Apple Silicon and Intel) and Windows (x64), released as open source under the MIT license. In this mode your ideas stay on your device. That is the core privacy promise: nothing is routed through Kompas servers just to capture, connect, or write.
Because the code is open source under MIT, you or a technical contact can inspect how data is stored and what, if anything, leaves the machine.
You choose the AI, including fully local models
Free mode does not ship its own AI. Instead you bring your own:
- An OpenAI or Anthropic API key
- A Claude Code or Codex subscription
- Local models run through Ollama
The Ollama option matters for privacy: if you run models locally, your notes and prompts do not need to leave your device for AI assistance at all. Using a hosted API key or subscription means those requests go to that provider under their terms, so the privacy boundary shifts to whichever service you connect.
Pro cloud: sync is the real consideration
Pro is the cloud plan, offered with a 7-day free trial and no credit card required. It adds two things the free desktop does not: cross-device sync between your desktops, and frontier AI built in with no setup or API keys, plus included usage with clear limits and optional top-ups.
The privacy question here is straightforward: syncing means your sessions are stored beyond a single device. If your notes are sensitive, decide whether you are comfortable with that before moving to Pro. The free desktop remains the option that keeps everything local.
What Kompas states about your data
The site's own wording is that your ideas are "never sold, never used to train models. Just yours." Treat that as the vendor's stated policy rather than something you can verify from the outside; the open-source desktop code is the part you can actually check.
A practical way to decide
| Your priority | Better fit | Why |
|---|---|---|
| Maximum privacy, notes stay on one machine | Free desktop | Local-first storage; no sync |
| No data leaving the device at all, even for AI | Free desktop + Ollama | Local models keep prompts on-device |
| Using your existing OpenAI/Anthropic key or subscription | Free desktop | Bring-your-own AI |
| Working across multiple desktops | Pro cloud | Cross-device sync (data leaves the single device) |
| Built-in AI with no key management | Pro cloud | Frontier AI included, with usage limits |
If privacy is your deciding factor, start with the free desktop and only move to Pro once you actually need sync. That way you keep your notes local until a concrete need justifies the cloud step.