
The user is on hotel Wi-Fi. Or school Wi-Fi. Or the office guest SSID. WireGuard never finishes. The ring spins. They screenshot your logo and write does not work. You did not lose a protocol debate. You lost a sentence the app should have shown.
WireGuard, as people actually deploy it, rides UDP. Hotels, schools, and companies often allow the web and drop the rest. UDP 51820 is a common example of a port that never leaves the building. The handshake times out. That is the whole incident until your copy makes it worse.
This is not inherit-versus-own. That essay is which claims you get to print because the platform shipped WireGuard. This is not the hotel amenity playbook with QR codes and night auditors. This is not whether TCP 443 should be a homepage story. That is the OpenVPN TCP fallback piece. Here the job is narrower: the filter, the symptom, and what the branded client should say before a human types.
A VPN encrypts a hop. Wikipedia's VPN page will not write your timeout string. RFC 8446 is TLS 1.3. It will not open UDP for you.
Klox consumer apps speak WireGuard, OpenVPN, OpenConnect, and Shadowsocks. Five devices. Yearly from $2.83 a month. Seven-day money-back. I will not invent a third protocol so the fallback paragraph sounds fancier. Your white-label brand owns the error text. The network owns whether UDP is even allowed.
I want the app to be boringly specific. Could not complete WireGuard handshake. Try OpenVPN. If the client has TCP, say that too, with the same words as settings. Cute copy is how people think the internet is broken. Scary copy is how they think you are a scareware company. Pick the boring middle.
Related reading: White-Label VPN and Waf and White-Label VPN and Wagtail Admin. White-Label VPN and Wagtail Cloud and White-Label VPN and Wagtail on a branded site. What is a VPN? and Download KloxVPN.
Looking for a reliable VPN?
KloxVPN — from $2.83/month. Apps for every device.
The handshake that never starts
A blocked UDP path does not always look like an error. It looks like Connecting for 20 seconds, then a generic fail, then a retry that fails the same way. Users call that the app is broken. Engineers call that a timeout. Brands need a third name: this network is not passing WireGuard. Put that in the UI or you will spend the year translating.
Do not lead with your Wi-Fi is unsafe. They already paid for a tunnel. They need a next step. Unsafe is a homepage scare. It is the wrong noun at minute two on a lobby SSID.
If the client retries WireGuard forever with no hint, you trained them to wait. Waiting is how they miss dinner and leave a one-star. I would rather an early, slightly ugly banner than a beautiful spinner.
- 1Skim the seating / order diagram.
- 2Do the numbered steps once on your real network.
- 3Use the FAQ if a sentence was too long.
- 4Follow one related article — not ten tabs.
How to read this page
| What they see | Likely cause | In-app sentence | Next step |
|---|---|---|---|
| Connecting, then timeout | UDP to WireGuard dropped | Could not complete WireGuard handshake on this network | Try OpenVPN |
| No internet, splash page never loads | Captive portal not finished | This Wi-Fi needs a sign-in page first. Pause VPN, join, then connect | Portal, then tunnel |
| Connects on LTE, fails on Wi-Fi | This AP filters UDP | Mobile data worked. This Wi-Fi may block WireGuard | OpenVPN on Wi-Fi |
| Works on home, fails at school | School filter | This network may block VPN UDP | OpenVPN TCP if you ship it |
| Works at home, fails in office | Corporate filter or inspection | This network may block or inspect VPN | OpenVPN TCP; do not promise stealth |
Handshake timed out is a network fact. Your logo makes it a brand fact.
— KloxVPN operator notes
Cloudflare Learning: What is a VPN?
Timeout is not a speed complaint
If they never connect, do not send the slow streaming macro. Wrong folder. Timeout is transports and filters. Slow after connect is a different day.
Do not blame their phone first
Reinstall is the last step, not the first. A filter does not care that you cleared cache. Leading with reinstall is how you sound like you do not understand Wi-Fi.
Three filters: hotel, school, office
They are not the same machine. A hotel wants a splash page, a payment, maybe a filter that treats UDP like abuse. A school wants minors off the open web and often ships a vendor that blocks VPN by category. An office wants data loss control and sometimes TLS inspection. One macro for all three is how you look lost.
Hotel: captive portal first. Then UDP. Then OpenVPN. The amenity article is if you are selling to the hotel. This article is if your consumer user is a guest on someone else's lobby SSID.
School: they may see a page that names VPN. Your copy should not dare them to jailbreak the filter. Offer OpenVPN. If it still dies, tell the truth: this network may not allow VPNs. Refund policy is a later conversation. Lying that you always work on campus is how you buy chargebacks.
Office: guest Wi-Fi might be hotel-like. Corporate SSID might inspect. Do not print we beat inspection. Print try OpenVPN TCP. If the company forbids VPN, that is an HR problem wearing your icon.
Airports and cafes
Treat them like hotels until proven otherwise. Portal, then UDP, then TCP. Do not write a special airport protocol. You do not have one.
Home ISP is the control group
If WireGuard works at home and fails on the trip, the app is probably fine. Say that. It calms people. It also stops you from shipping a fake app bug every Monday after a long weekend.
Captive portal versus UDP drop
Two failures get mashed into VPN broken. One: they never accepted the portal, so nothing including WireGuard can reach the internet. Two: they are online, Chrome loads, WireGuard still times out. If your string does not distinguish those, agents will not either.
Sequence for hotels: pause or disconnect the VPN, complete the portal, then connect. Kill switch copy can fight this. If fail-closed is on, the portal never loads. Point at the kill-switch article. Do not pretend transports fix a switch that is strangling DHCP.
Once the portal is done, Chrome works, and WireGuard still dies, you are in UDP-drop land. Now OpenVPN. Now TCP if you ship it. Now a sentence that does not mention the portal again.
Detecting portal is messy
Some clients detect captive portals. Some do not. Do not promise we detect hotel login if the build does not. A wrong detection string is worse than a generic timeout.
TVs and consoles
A Fire Stick on hotel Wi-Fi is a different support object. If your five-device plan includes a TV, write a TV paragraph. Protocol pickers on TVs are often worse. Tell the truth about what that client can switch.
What the app should say
Write the timeout like a human who has been in a lobby. Could not connect with WireGuard. This Wi-Fi may block it. Try OpenVPN in settings. Button: Switch to OpenVPN. Secondary: View steps. Do not add a lecture about encryption. They already bought encryption. They need a tap.
If you cannot deep-link to the protocol picker, the button is a liar. Ship a picker path or drop the button. A button that opens Chat is an escape hatch. Fine if you staff it. Not fine if the hotel is at midnight in another timezone.
Match settings labels. If settings say OpenVPN (TCP) and the banner says stealth, you created a scavenger hunt. The TCP-fallback article is how you decide whether TCP is a product. This article is the banner that points at whatever you decided.
Tone
No panic. No dark web. No your ISP can see you in this hotel. The tunnel is down. Give them the next protocol. Save the threat model for the homepage.
Localization
Translate the next step, not a poem. Handshake, WireGuard, OpenVPN stay as names. Blocked, try, settings are the words that must survive a bad translation.
Fallback order the user can follow
A printed ladder beats a smart Auto nobody understands. 1. Finish the portal. 2. WireGuard. 3. OpenVPN UDP if you expose it. 4. OpenVPN TCP if you ship it. 5. Mobile data as a test. 6. Stop, and tell them this network may not allow a VPN.
Auto can climb that ladder if you still show which rung won. A silent Auto that lands on TCP while the UI still says WireGuard is how tickets say your app lies. Show the active protocol on the home card. Small type. Always.
Klox consumer is WireGuard plus OpenVPN. The ladder ends there in this post. I will not add a mystery protocol for flavor.
Mobile data as a test
If LTE works and the SSID fails, you have a Wi-Fi policy problem. Tell them that. It prevents the reinstall death spiral. Watch data caps. Do not leave them on LTE as the only advice.
Remember last-good per network if you have it
Some clients remember. Confirm before you advertise it. Promising we remember hotel Wi-Fi when you do not is a second timeout with extra disappointment.
Help center versus the glass versus the agent
Three layers. The glass is for the person who will never search. The help center is for the person who will. The agent is for the leftover. If the glass is mute, the other two drown.
Help article title I actually want: VPN will not connect on hotel or school Wi-Fi. Not troubleshooting connectivity issues. The first title is searchable. The second is a folder name.
Agent macro: same ladder, same nouns as the banner. Attach screenshots per OS. Five devices means they may be on a laptop while chatting from a phone. Ask which device is failing. I still see macros that assume Android.
Do not send a WireGuard explainer
The inherit article and the consumer WireGuard primer exist. A timeout ticket is not a teaching moment about Noise protocol. Save the lecture.
Competitor help centers
NordVPN's public help is a specimen of how a large consumer brand talks about protocol switching. Read it. Do not paste it. Your strings, your picker, your hours.
Corporate inspection and VPN detected
Sometimes the office page says VPN detected and kills you. Sometimes inspection breaks the handshake in a way that looks like a timeout. You will not tell those apart from a banner. Do not claim you will.
Allowed sentence: This network may block or inspect VPNs. Try OpenVPN TCP. If it still fails, the network may not allow VPNs. Disallowed: we disguise traffic as HTTPS so IT cannot see you. That is a fight with their employer wearing your logo. Decline that fight in writing.
RFC 8446 still is not your camouflage. Port 443 still is not invisibility. The TCP-fallback post is the honesty paragraph. Repeat a short version here so an agent does not invent stealth in chat.
Managed laptops
If the device is company-owned, a consumer VPN brand is often against policy. Your AUP and the abuse-desk article are related. The timeout string should not coach them to evade MDM.
Guest SSID versus corporate SSID
Ask which. Guest is often hotel-like. Corporate is often policy. One screenshot of the Wi-Fi name saves ten minutes.
School pages that name VPN
A filter that shows VPN blocked is doing you a favor. You can stop guessing. Your reply is not a bypass tutorial. Your reply is we may not work on this network, here is OpenVPN anyway, here is the refund window if you just bought.
Minors plus VPN advice is how you get a conversation you do not want. Keep the steps technical and short. Do not joke about sneaking past the school. Cute tweets become exhibits.
Campus as a product is the university white-label post. This is the consumer user who happens to be a student. Different buyer. Same UDP.
Library Wi-Fi
Treat like school until proven kinder. Public libraries often share filter vendors. Same ladder. Same refusal to coach evasion.
Do not invent city counts in the school paragraph
Servers in more cities will not open UDP 51820 on a Cisco that drops it. Geography is the wrong comfort.
Five devices, five networks
The account is one. The paths are many. A router at home on WireGuard, a phone at the hotel on a timeout, a tablet at school on a filter page. Support that treats the account as one device will give router advice to a phone.
Ask device, OS, Wi-Fi versus LTE, whether Chrome loads, whether they passed a splash page, which protocol the home card shows. That is the intake. Without it you are performing.
Klox consumer is five devices. Your brand may pick another number. The intake does not change.
Do not remote into their hotel
You will not. Give them the ladder. If they cannot climb it, refund or wait for better Wi-Fi. Pretending you can SSH their lobby AP is how you promise magic.
Shared family accounts
A parent chats. A kid is on the school tablet. The timeout is the kid's. The parent will describe home Wi-Fi. Ask who is failing. It sounds basic. It is the usual miss.
What you must not tell them
No always works on any Wi-Fi. No guaranteed hotel bypass. No city counts as a fix. No SOC 2. No we are undetectable. No turn off your school filter. No the protocol is broken so we will refund Mbps you never had.
You may say WireGuard uses UDP and some networks block it. You may say OpenVPN, including TCP, is the usual next try. You may say if both fail, this network may not allow VPNs. That is an adult paragraph. Use it on the site and in the app.
Do not blame WireGuard as a technology. Blame this path. The inherit article already told you not to become a protocol author in public. The timeout is not your chance to recant WireGuard on the homepage.
Reviews
Reply to the hotel timeout review with the ladder, not a debate. Invite them to try OpenVPN. If they already refunded, say so calmly. Do not argue that UDP is faster. They do not care. They wanted connected.
Ads
If you buy travel keywords, the timeout string has to exist before the campaign. Buying hotel VPN and shipping a mute spinner is paid refunds.
Test the string on a real AP
Before you freeze copy, fail on purpose. A cheap travel router that drops UDP is enough. A school-like filter if you can do it legally on your own lab. The office guest SSID on a Sunday. Watch the banner. Time how long until it appears. If it appears after they already left the app, it did not appear.
Have a stranger try without a briefing. If they cannot find OpenVPN in 30 seconds, the label is wrong. You are too close to the settings tree.
Then freeze a weekly count: handshake timeout tickets, switched to OpenVPN, refunds that mention hotel or school. If those counts are not boring, your homepage can still lead with WireGuard. Your timeout string cannot stay mute.
QA on each OS
iOS, Android, Windows, macOS if you ship them. The banner will not match. Screenshot each. Put the PNGs in the help article the same week you ship the string.
Seven-day money-back
Klox consumer is seven days on first purchase. Your brand may copy that or not. A timeout with no next step inside that window is a refund you earned. TCP that only exists in a weekday macro does not count as a next step at 23:40.
What is not a UDP block
Agents will call everything a hotel block. Wrong. Auth expired. Account past due. Kill switch eating the portal. IPv6 leftover. DNS on the phone still pointing at the ISP. A burned exit the user already connected to last week. If you teach the inbox that timeout equals UDP, you will miss the cases where WireGuard is fine and billing is not.
Intake first: are they signed in? Does the home card show a plan? Did they just fail a card? Then portal. Then UDP. The dunning article is past-due access. The kill-switch article is fail-closed. This article starts after those are ruled out.
I would put a six-line checklist at the top of the timeout macro. If agents skip it, they will switch a paying home-fiber user to TCP for no reason and then you will read a slow app review. TCP is a tool. It is not a personality.
Account paused looks like a timeout
Some clients show Connecting when the peer was revoked. That is not a filter. That is entitlement. If refunds and dunning kill access, the timeout string must not send those users into OpenVPN. Send them to billing.
Wrong location versus wrong transport
A far exit can feel like a hang on a bad AP. Still not UDP drop. If Chrome loads and the handshake dies fast, think filter. If everything is slow including the web, think the hotel AP is junk. Different sentence.
Publish a help URL the banner can use
The in-app sentence should link to a stable URL you own: /help/restricted-wifi or whatever your help center actually serves. Not a Google Doc. Not a Notion page that dies when a contractor leaves. The banner is useless if View steps 404s.
That article should be the ladder, screenshots, and the honest end: this network may not allow VPNs. Date it. When you rename OpenVPN TCP in the app, edit the article the same day. Stale help is how people say your brand is fake.
If you do not have a help center yet, the knowledge-base post is the staffing of articles. Still ship the banner. A banner without a URL beats a URL without a banner. Both is the adult version.
Deep links
If the client can open settings at the protocol picker, use that. A web article that says tap the slider is a last resort for a native app. Confirm deep links per OS. iOS will break them after a rename.
Do not put this on the homepage
The homepage can mention travel. It should not lead with we sometimes fail on UDP. The timeout UI is the right place for the failure. Homepage honesty is WireGuard, OpenVPN, OpenConnect, and Shadowsocks exist. Failure honesty is the banner.
Measure the timeout or you will argue from vibes
Count tickets whose first line is will not connect on Wi-Fi. Tag hotel, school, office, other. Tag whether they had already finished a portal. Tag whether they switched to OpenVPN. Tag refunds inside seven days that mention connect. If you cannot tag, you will relaunch the same banner next year.
I do not want a fake dashboard percentage. I want a weekly number a founder can stand. If hotel tags rise after a travel campaign, pause the campaign until the banner ships. Buying traffic into a mute spinner is paid refunds. Klox consumer seven-day money-back will happily take that money back out. Your brand's window will too if you copied it.
Share the count with whoever owns store screenshots. A sofa Connected shot next to a travel ad is the mismatch users feel in the lobby.
Do not optimize for zero timeouts
UDP will die somewhere. Zero is a lie. Optimize for time-to-OpenVPN and for the honest end state. That is an ops target, not a protocol miracle.
White-label versus reseller
If you are a reseller of Klox the consumer product, your users may see Klox copy. If you are white-label, they see yours. Do not paste a Klox timeout string into a brand that renamed the picker. Nouns have to match the glass they hold.
Key Takeaways
When WireGuard UDP is blocked, the branded app either names the failure and the next protocol, or it becomes a spinner with your logo on it. Hotels, schools, and offices are different filters. Captive portal is a different failure from a UDP drop. Mix those and the inbox will mix them too.
Klox ships WireGuard, OpenVPN, OpenConnect, and Shadowsocks. Five devices. From $2.83 a month on the consumer yearly. Seven-day money-back on first purchase for Klox consumer. Your white-label brand still writes the timeout sentence. Do not invent extra protocols here. Do not invent stealth. Do not invent city counts that open a closed port.
If you want apps where you own that sentence, start from white-label. Then fail WireGuard on a real AP before you buy a travel ad.
Related Resources
Own the timeout sentence on your apps
White-label puts your brand on clients that speak WireGuard, OpenVPN, OpenConnect, and Shadowsocks. You still write the handshake-timeout copy and the fallback order. Klox consumer remains from $2.83 a month, five devices, seven-day money-back — that SKU is not your hotel banner.
See white-label VPNFrequently Asked Questions
KloxVPN Team
Experts in VPN infrastructure, network security, and online privacy. The KloxVPN team has been building and operating VPN services since 2019, providing consumer and white-label VPN solutions to thousands of users worldwide.