A VPN protects data in transit on a specific part of the network, but it requires trust in whoever runs the server and the app. The risks come from opaque data collection, misconfiguration and malware. Let’s look at whether it’s safe to use a VPN, what limitations the technology has and how to check a service before connecting.
Is it safe to use a VPN
A modern VPN creates an encrypted tunnel between your device and a VPN server. Traffic inside it is harder for outsiders on the local network and along the path to that server to read or tamper with. When you go online through a VPN, websites see the VPN node’s address instead of your original IP. These are useful properties, but their scope is limited by routing and by how well the client is implemented.
A VPN doesn’t check whether a website is genuine, doesn’t fix an infected system and doesn’t stop you from sending your password to a scammer. If malware reads data before it’s encrypted, the tunnel won’t save it. Account sign-ins, cookies and your browser fingerprint also remain. So whether a VPN is safe for a particular person depends on the threat: hiding your home IP and protecting yourself from phishing are different tasks.
Beyond the VPN server, content is protected by the application’s protocol. For websites, that’s HTTPS: it encrypts the link between the browser and the site and verifies the certificate. A VPN doesn’t turn plain HTTP into HTTPS. Cloudflare’s explainer covers the role of HTTPS; Mozilla shows how to assess the state of a connection in the browser.
A strong protocol doesn’t guarantee the server owner is honest. A VPN shifts part of the trust from your access provider to another intermediary. So check the protocol, the settings and the data policy.
How a VPN can be dangerous
Collecting and sharing user data
A service may process your connecting address, session times, device information and account data. Some apps include analytics and advertising libraries. The risk arises when what is collected, how long it’s kept and who receives it are unclear or don’t match the promises made. Being technically able to see data and actually keeping logs aren’t the same thing, but you need to find out about both.
Find the privacy policy: are IP addresses, diagnostics and connection history stored? Check whether you can delete your account and control optional analytics. A general promise of privacy isn’t enough.
IP address and DNS leaks
A leak usually means data leaving outside the expected protected route. For example, IPv4 goes through the VPN while IPv6 stays direct; DNS is handled by your ISP’s resolver; after a drop, the client keeps working without the tunnel. With split tunnelling, some direct connections may be intentional. What matters is comparing the result with the setting you chose, not expecting every client to behave the same way.
Check your public IPv4 and IPv6, the DNS servers in use and the behaviour after reconnecting. Then repeat the check when switching from Wi-Fi to mobile data. If there’s a kill switch, find out what it covers: traffic blocking has to work on your platform and in your connection mode. The mere presence of the word in a feature list doesn’t confirm the result.
Malicious VPN apps and data theft
Apps with malicious code or excessive permissions can be distributed under the guise of a VPN. A random APK, a “cracked subscription” or a copy of a well-known brand are reasons to stop and check the source. Access to contacts, SMS or accessibility features should have a clear justification. The system permission to create a VPN connection is normal in itself, but it doesn’t make the publisher trustworthy.
Be especially wary of being asked to install a root certificate for an ordinary personal VPN. Such a certificate can let an intermediary intercept HTTPS if configured to do so. Don’t ignore browser certificate warnings. If you suspect something, remove the app and any unnecessary profile, scan the system and, from a trusted device, change any passwords that may have been exposed.
Why free VPN services can be dangerous
A free plan may be funded by sales of paid subscriptions and isn’t necessarily dangerous. A paid VPN doesn’t guarantee security either. Find out how the service is funded and the terms of your specific plan.
An opaque service may monetise advertising, analytics or the sharing of information about users. Advertising in itself isn’t data theft, but built-in trackers and unknown recipients increase the amount of information you’re trusting them with. Check whether data is used for ad profiling and whether you can opt out of optional collection. Granting permission to connect to a VPN shouldn’t replace an informed choice of terms.
If there’s no clear owner, no data-processing terms and no explanation of the business model, it’s better not to route sensitive accounts through such a service. Look into the specific product rather than generalising about all free VPNs. The FTC recommends checking a VPN app’s promises, its permissions and how the service is paid for, including possible monetisation of data.
What data the VPN server owner can see
For the connection to work, the VPN node receives the client’s original IP and processes the traffic. Connection time, duration and data volume are technically available. The server sees destination addresses; domain names may be visible through DNS and connection metadata. The exact set depends on the protocols, DNS encryption and the architecture. If you’re signed in to your account, the service also has your account information separately.
With ordinary HTTPS and no certificate substitution, a VPN provider doesn’t get your plaintext password, form contents or the full page path just because your traffic passes through its server. But HTTPS doesn’t hide all metadata. With unencrypted HTTP, the content is visible to an observer at the VPN exit. Tunnel encryption and application encryption complement each other and protect different legs of the path.
No-logs is a claim about limiting logging. It needs clarification: does it refer to browsing history, source IPs, connection times or only traffic content? Account, payment and technical information may be processed separately. It doesn’t mean there is no data about the user at all.
Check the policy, retention periods and the scope of the audit, if one has been published. The date, the system that was audited and an accessible report all matter: a past audit doesn’t guarantee today’s app is unchanged. The EFF also advises assessing how far you can trust the provider and the real limits of VPN protection.
How to choose a safe VPN and reduce the risks
- Define the task. Protecting your route on Wi-Fi, accessing a work network and changing your exit IP each require checking different properties. Don’t buy a promise of total anonymity instead of a specific feature.
- Check the provider. Look for the owner, contact details, a data policy and an update history. Reviews are useful as a signal, but an advertising ranking doesn’t replace technical documentation.
- Choose a supported protocol. WireGuard and OpenVPN are examples of open solutions. Their names don’t guarantee correct configuration, protection of all routes or an honest operator.
- Download from a trusted source. Go to the app store from the official website and check the publisher. For an installer outside the store, use the official channel and any available signature verification.
- Test the client on your own device. Make sure IPv4, IPv6 and DNS are covered as intended, and check what happens when the connection drops. Repeat the check after updates and settings changes.
MeduzaVPN supports WireGuard and OpenVPN; PERSONAL uses a personal VPS with a dedicated IP. The official security page also describes isolated protocol containers and the policy of not storing traffic content. Not storing content doesn’t mean no data is processed at all: check the policy and the terms.
Containers separate components, but they don’t replace device protection. A dedicated IP doesn’t make a user anonymous. Check direct-traffic blocking features in your chosen client, separately for each platform.
Keep your system updated, use HTTPS, unique passwords and multi-factor authentication. If the connection behaves unexpectedly, use our step-by-step troubleshooting and contact support, without sharing secret keys.
