No-Logs When the Brand Isn't the Network Operator

When a privacy page says we, users hear the brand they paid. They do not hear a processor annex. If you are not ready for that, do not use we for node operations. Use named roles. 'Your account data is held by [brand]. The VPN network is operated by [provider] under a processing agreement.' It is uglier than a slogan. It is also the sentence that keeps you out of a trap.

I would rather lose a few conversions than win them with a we that collapses in email one.

Hosted: the platform operates servers. You operate a brand, billing, and usually the controller role for customer accounts. Source: you can operate servers. You still might use a cloud host, which is another processor. Source is not automatic moral purity. It is more ssh and more homework.

Do not write hosted copy that sounds like source. Do not write source copy that ignores the cloud provider you actually used. The user does not care about your invoice layout. They care who could have seen a source IP.

Email, password hashes, device count, invoices, refunds, support tickets. That pile exists even if the nodes forget every handshake. No-logs marketing that pretends you have no personal data is how you get a reviewer who opens your password-reset form and laughs.

KloxVPN consumer plans include 5 devices and billing from $2.83/month on the yearly plan. That billing relationship is personal data. White-label brands have their own processors (often Stripe). Name them. 'We don't log' next to a Stripe checkout without a payment section is a policy hole, not a flex.

If you cannot set retention on the operator's nodes, do not publish a retention number you invented. If you cannot attest to disk wiping, do not write 'RAM-only servers' because it sounded good in a competitor FAQ. If you cannot see a warrant, do not promise how you would fight one.

You can promise how you handle account deletion in your portal. You can promise you will not build a dashboard that shows user browsing. You cannot promise a packet log you have never inspected does not exist. Inspect the contract. Ask for the operator's privacy notice. Link it. Dual notices are allowed. Dual fantasies are not.

If agents ask for destination URLs 'to debug streaming,' you are requesting usage data even if nodes do not store it. If agents can see last-connected timestamps in admin, that is a log, even if you call it telemetry. Align the admin panel, the ticket macros, and the privacy page. The mismatch is how a Reddit user proves you wrong with a screenshot of your own dashboard.

Ask the platform what the admin actually shows. User online now? Last handshake? Bytes? Source IP? If you do not know, you cannot write a policy.

Someone will use the network badly. Someone will send a legal request. The request will go to whichever entity looks like the operator, or to both. Your policy should say where requests should go, what you can actually produce (usually account and billing, not browsing), and that you will not invent records.

If the operator produces technical data you never saw, your public 'we had nothing' needs a footnote. 'The brand did not have traffic logs' is not the same as 'nobody did.' Precision is not optional once lawyers are in the thread.

Read Mullvad's no-logging policy as a specimen of tone and structure: they describe what they do not store, they write as the operator of the service, and they do not hide behind a slogan with no nouns. That is the bar for a network operator. It is not a template for a brand that does not operate the network, and it is not a claim about KloxVPN.

If you are the operator (source deployment, your nodes), you still should not clone another company's page. Your disk layout is not theirs. Your payment stack is not theirs. Steal the honesty, not the sentences.

A no-logs VPN landing page with a full advertising pixel package is a joke people will tell without you. If you need ads, isolate them, disclose them, and do not run them inside the authenticated portal if you can help it. If you cannot help it, say so.

I would rather a quieter homepage than a privacy theater that loads six third parties before the first paragraph. Users who care will look. Reviewers who care will look. Your competitor's affiliate will look with a recorder on.

Data-safety forms that say no data collected while you require an email login will get you in trouble. Match the form to the portal. If you collect crash reports, say so. If you collect none, make sure the SDK list is empty, not just the prose.

White-label means the binary might include libraries you did not personally choose. Ask for the list. Your privacy nutrition label is yours even if the module was inherited.

Agents should not say 'we can look up your sessions' if you claim you cannot. They should not say 'we never keep emails' when the ticket system is the email. Train them on the same nouns as the policy: account data vs traffic data. If they cannot keep those nouns straight, they should not work tickets yet.

If an affiliate writes 'this brand never logs and is based in [fantasy country],' and you pay them, you will own the screenshot. Give affiliates a claims sheet. Ban the rest. VPN affiliate culture will invent audits for you if you do not police it.

I would claw back commission for banned claims. Soft feelings about partners are how false no-logs copy spreads faster than your legal page can load.

You can say your apps include WireGuard. You cannot say you are 'more no-logs' than a company that published a detailed operator policy and a long public track record unless you have an equivalent operator story. Compete on brand, niche, language, support hours, price. Do not compete on a court case you did not have.

A list of device names for a 5-device plan is account data. A list of every IP those devices used as source addresses for 90 days is closer to connection logging. Know which one your admin shows. If you show last IP, say so. If you do not know, you are writing fiction.

Parents will ticket you to kick a device. You need enough account data to do that. That is not a betrayal of no-logs. It is how a family plan works. Explain it without mixing in 'we watch your traffic.'

Users will send screenshots of URLs. That is usage-ish data in your mailbox. Retain tickets on a schedule. Do not promise you never see destinations if your streaming macro asks for the exact show title and URL. Change the macro or change the promise.

Usually: whether an email has an account, subscription dates, maybe invoice metadata, maybe support transcripts. Usually not: a browsing diary you claimed not to have. If a request asks for the diary and you do not have it, say you do not have it. Do not stall in a way that looks like you are searching a system that does not exist.

If the operator might have technical records, do not answer for them. Route. Document that you routed. Your future self will need that file.

A journalist will ask if you log. If you hesitate, they will write the hesitation. Have a paragraph approved before launch. Not a slogan. A paragraph with roles, links, and what you will not claim. Put the same paragraph in the press kit and the help center so they cannot be played against each other.

When you leave a platform, account data should come with you or be deleted on a written schedule. Node-side data, if any, should follow the operator's policy, not your feelings. Write both into the customer-facing deletion section so you do not promise a wipe you cannot trigger.

This is unglamorous. It is also the difference between a brand and a landing page.

Put a clause that the brand will not attribute operator claims the operator did not make. Then obey it. The fastest way to poison a white-label relationship is a partner ad that invents an audit for both of you.

Date it. When you add a chat tool, update it. When you enable a new protocol, you probably do not need a new privacy novel, but when you add a crash reporter you do. Store listings should match the dated page.

KloxVPN's consumer privacy page lives at /privacy. A white-label brand needs its own URL on its own domain. Linking ours as if it were yours is a controller error. Do not do it.

Good is specific nouns, named processors, separate columns for account vs traffic, no stolen raid stories, no SOC 2 badge you do not have, no audit PDF you cannot produce, and support macros that use the same nouns. Good is also a 7-day money-back you can execute if you offer one, because refunds are part of trust even if they are not logs.

Perfect is not available. Specific is.