Softplorer Logo

exposure vs control

VPN for Security

A VPN doesn't make you secure. It closes one specific gap — and leaves the others exactly where they were.

You came here because: I want a kill switch so nothing leaks

This fits you if

  • You connect to public networks regularly and don't always remember to activate
  • You want to know the protection is real, not just claimed
  • Your threat model goes beyond a misconfigured café network

What's happening

You're on a network you didn't set up. Someone else configured the router, someone else controls the traffic, and you have no visibility into either. You connect anyway — because you need to, because it's convenient, because the alternative is not working at all. In that moment, everything you send is moving through infrastructure you can't inspect and don't trust.

A VPN on an untrusted network encrypts your traffic before it leaves your device. The network sees a connection to a VPN endpoint — not your DNS queries, not the sites you visit, not the content of what you're doing. That's a meaningful protection against passive observation on a misconfigured or hostile network. It's not a meaningful protection against most of the other ways things go wrong.

The security use case is different from privacy or streaming in one important way: the moment of risk is acute, not chronic. You're not worried about your ISP building a profile over months. You're on an airport network right now, and you don't know who's on it with you. The providers that serve this well are the ones that activate reliably and completely — not the ones with the most features.

Philosophies

NordVPN

Scale done reliably

View breakdown

Auto-connect on untrusted networks closes the gap that matters most for security use: the window between joining a network and remembering to activate the VPN. Threat Protection adds DNS-level filtering that catches some risks before the tunnel is even established. The architecture isn't open for inspection, so you're trusting the outcome — but the outcome is consistent and the activation is automatic, which is what security use cases actually require.

NordVPNVisit NordVPN
ExpressVPN

Complexity should be invisible

View breakdown

Network Lock holds all traffic if the VPN tunnel drops — which matters in the seconds between a network change and a successful reconnect. The activation experience is immediate and the interface doesn't require decisions mid-connection. Users who want to understand the mechanism rather than trust the result will find the architecture opaque; users who want protection that works without thinking about it will find it does exactly that.

ExpressVPNVisit ExpressVPN
ProtonVPN

Verification over convenience

View breakdown

For users whose threat model extends beyond passive network observation — journalists, activists, anyone operating in environments where the attacker is sophisticated — Proton's verifiable architecture changes the calculation. Open-source apps and independent audits mean the protection can be confirmed rather than assumed. The trade-off is setup friction: the product requires more attention than alternatives that handle everything automatically.

ProtonVPNVisit ProtonVPN
Mullvad

Identity should not be required

View breakdown

No account and no email means there's less to compromise even if the provider is targeted. On an untrusted network, this structural approach to identity minimization adds a layer that policy-based privacy claims don't. The ecosystem is narrow — fewer apps, less device flexibility — and the setup assumes technical familiarity that casual users connecting to airport Wi-Fi typically don't have.

MullvadVisit Mullvad

Recognize yourself

You connect to public networks regularly and don't always remember to activate

Manual activation relies on a habit that travel and distraction reliably break. The gap between connecting to a network and turning the VPN on is exactly when passive exposure happens. Providers without auto-connect on untrusted networks leave that gap open every time — and you'll close it consistently right up until the day you don't.

You want to know the protection is real, not just claimed

Kill switch implementations vary more than the marketing suggests. Some hold all traffic. Some hold only browser traffic. Some have edge cases around network changes that let packets through before the tunnel re-establishes. The only way to know which category your provider falls into is to test it — or choose one whose code is open enough that someone already has.

Your threat model goes beyond a misconfigured café network

Commercial VPNs are built for the mass market. Sophisticated adversaries with resources to target individuals specifically operate in a different category than passive network sniffing. A consumer VPN changes your IP and encrypts your traffic — it doesn't make you invisible to a motivated, well-resourced attacker. Providers built for verifiability narrow the gap more than those built for convenience, but neither closes it entirely.

You're handling sensitive work on networks you don't control

The kill switch behavior under load matters here — not just whether it activates, but whether it holds through a network transition mid-upload, mid-call, or mid-sync. Providers that handle this gracefully do so through protocol design. Providers that don't will expose your traffic in exactly the moments when you're least able to notice.

No guarantees

A VPN encrypts the connection between your device and the VPN server. It doesn't encrypt what happens after — the server to the destination, the destination itself, or anything involving an account you're logged into. Logging into a personal account through a VPN doesn't make that account's activity private from the service hosting it.

Most public network risks are mundane. Passive traffic observation on a misconfigured network is more common than a targeted attack. A VPN addresses both — but the protection is proportionate to the threat. Calibrating your setup to your actual exposure rather than a worst-case scenario is reasonable.

A VPN is one layer. Unpatched software, reused passwords, phishing, and compromised devices create vulnerabilities that encrypted traffic doesn't touch. Security that begins and ends with a VPN has a significant surface area left unaddressed.

Where to go next