The Speed Test Passed—But Only One VPN Kept My Video Call Alive in Egypt

“We can see you,” the client said. “We just can’t hear you.”

I was sitting at the end of a hotel bed in Cairo with my laptop balanced on a suitcase because the desk lamp had claimed the only convenient socket. The hotel Wi-Fi had passed a speed test at 48 Mbps. My slides had uploaded. Email worked. Yet the meeting showed my microphone as connected while every sentence arrived as silence. I left the call, restarted the laptop and joined again. This time, the audio lasted twelve seconds before the screen froze.

The presentation was mine to lead.

Thirty-two people were waiting.

My colleague sent a private message:

“Do you have a backup?”

I thought I did.

Fast internet was not the same as a usable call

The hotel connection looked healthy whenever I used it for ordinary browsing.

News pages opened quickly. A large PDF downloaded. The meeting application showed full signal bars. None of that explained why my voice kept disappearing.

A website can pause briefly and continue loading. A live conversation cannot lose several seconds and still feel live.

Egypt added another complication. WhatsApp messaging remains available, but internet calling has faced restrictions and throttling. (Reuters) Other voice and video platforms can also behave differently depending on the network and the route carrying the call.

That creates a familiar traveller’s puzzle:

Messages work.

Photographs send.

Websites load.

The call refuses to begin—or begins and collapses a minute later.

The speed test had measured how quickly the hotel could move a short burst of data. It had not shown whether the connection could carry a continuous meeting.

My colleague messaged again:

“We’re filling time. Maybe five minutes.”

I opened the commercial VPN I normally used while travelling.

A long server list gave me more restarts

The established provider had years of public history, a polished application and more server locations than I would ever need.

I selected the nearest recommended route.

The VPN connected, and I rejoined the meeting.

My audio returned.

I apologised, shared my screen and reached the third slide before the client interrupted.

“You’re breaking up again.”

The video became a row of blurred squares. The meeting application reduced the image quality, removed my camera and then disconnected the call.

I tried another country.

Then another.

Each server gave me a fresh route, but none kept the meeting alive. One remained on Connecting until I cancelled it. Another connected while leaving the meeting page unable to load.

Egyptian networks have a documented history of interfering with VPN and privacy services. OONI measurements have recorded blocking involving VPN-related websites and services. (Ooni) In practical terms, the provider’s connection pattern, login page or support site can become part of the same problem the VPN is supposed to solve.

Users describe the result more simply: a VPN works for a few minutes, stops, and sends them back through a list of alternatives while the call remains unusable. (Reddit)

I had now spent more time changing countries than speaking to the client.

The large server list was not helping me finish the presentation.

It was giving me more ways to restart it.

Mobile data solved only half the problem

I switched off the hotel Wi-Fi and enabled the hotspot on my phone.

The laptop moved to mobile data.

Without the VPN, the meeting opened again. The client could hear me, although the audio sounded thin and delayed.

I returned to the slides.

For almost two minutes, the call held.

Then the mobile signal fell from four bars to two. My screen share stopped moving. The meeting application displayed Poor network connection and disconnected.

I walked toward the window while trying to rejoin.

The signal improved.

The call returned.

Then the laptop noticed the hotel Wi-Fi again and shifted back. The connection changed underneath the meeting, and I disappeared for a second time.

The pattern was finally clear.

I had one problem getting the video call through the local network and another keeping it alive when the laptop moved between hotel Wi-Fi and mobile data.

Choosing another country addressed neither one.

I needed a protected connection built for the situation in front of me.

The smaller app began with the problem

I closed the established provider and opened the smaller backup already installed on the laptop.

It did not begin with a map.

The first screen asked what kind of network I was using. I selected the preset for restrictive and unstable connections.

The VPN connected.

I returned to the meeting link.

The waiting room opened immediately. My colleague admitted me, and the client’s face appeared in the main window.

“Can you hear me now?” I asked.

“Yes. Perfectly.”

I shared the presentation again.

Slide three.

Slide four.

The hotel Wi-Fi weakened as someone nearby began streaming. Instead of waiting for the call to collapse, I switched the laptop to my phone hotspot.

The video paused for less than a second.

My voice continued.

I reached the demonstration section and opened the client’s live dashboard. The page loaded inside the screen share. No one asked me to repeat the explanation.

At slide nine, the mobile signal weakened near the window. I returned to hotel Wi-Fi.

The call remained open.

The client asked a question about the final chart. I answered it, moved to the appendix and completed the presentation.

Twenty-four minutes after the meeting had begun, my colleague wrote:

“Audio is still clean.”

That was the first useful status message I had received all morning.

The smaller app combined traffic obfuscation with an HTTP/3-based connection. One helped the protected traffic avoid an obvious conventional VPN pattern; the other allowed the connection to recover when the laptop changed networks. (IETF)

The visible result was simpler.

The meeting survived both switches.

The call ended because we were finished

When I stopped sharing my screen, the client remained on the line.

We discussed the next delivery date, assigned two follow-up tasks and scheduled another meeting for the following week.

Then everyone said goodbye.

I clicked Leave.

For the first time that morning, the call ended because the conversation was over rather than because the network had removed me from it.

I could not observe the hotel provider’s internal filtering rules or identify exactly which part of the earlier VPN traffic had been disrupted. I could compare the outcomes: ordinary browsing passed every test, the established provider repeatedly lost the meeting, and the smaller app carried the same call across hotel Wi-Fi and mobile data without restarting it.

That changed what “reliable video calling” meant to me.

It was not the highest number on a speed test.

It was staying in the meeting when the network changed.

The second device joined without another login

After the call, the client asked me to send a photograph of a signed page stored on my phone.

The phone was still providing the hotspot, but it was not connected through the smaller service. I expected to search for another password, approve another login and risk disturbing the setup that had finally worked.

Instead, the laptop displayed a verification code.

I entered it on the phone.

The second device connected without another conventional email-and-password account. I opened the document scanner, captured the page and uploaded it to the client’s shared folder.

The file appeared on the laptop a few seconds later.

That was not what saved the presentation. The restrictive-network preset and stable handoffs had already solved the urgent problem.

The code simply removed the next interruption when the task moved to another device.

In a hotel room where every extra login depended on unreliable Wi-Fi, that was enough reason to keep the app installed.

A real backup must survive the switch

My original backup plan had been vague:

Use the hotel Wi-Fi.

Keep mobile data nearby.

Install a reputable VPN.

That sounded responsible until all three connections behaved differently during the same meeting.

The missing test was not another speed check. It was the handoff.

Could the exact meeting platform open?

Could the client hear me?

Could the call survive a move from Wi-Fi to mobile data?

Could the VPN connect without sending me through server experiments or an account-recovery process?

The established service had a mature interface and a much larger choice of locations. Its weakness appeared when the hotel network became unstable: every reconnection interrupted the active call.

The smaller service has fewer locations, fewer independent ratings and a shorter public history. That is its clearest limitation.

But the meeting did not need another flag on a map.

It needed one usable route that could pass through the network and remain alive when the laptop changed connections.

Reliability was measured in completed sentences

After the call, I ran the speed test again.

The hotel result was still close to 48 Mbps.

The number had not been false. It had simply answered the wrong question.

It showed that the connection could move data quickly for a moment.

It did not show whether the call would pass through the network, whether the VPN would remain usable, or whether the meeting would survive the laptop moving onto my phone.

The established provider gave me many servers and several fresh starts.

The smaller app gave me one continuous conversation.

In Egypt, the reliable video-call setup was the one that let me reach the final slide without disappearing between sentences.