Which VPN Actually Works in India Right Now? I Found Out During the Telegram Block

At 5:52 p.m., the link for my coding class was ready and Telegram was not. The desktop app showed Connecting… while eighty-three students waited for a 6 p.m. session. I blamed the coworking Wi-Fi, switched my laptop to a Jio hotspot and restarted Telegram. The group still would not open.

The class itself was hosted elsewhere. Telegram was where I had posted the timetable, exercises and last-minute instructions for more than a year. Without access to the group, students had no meeting link and no explanation for why I had disappeared.

That afternoon, the problem was larger than my laptop. India had temporarily blocked Telegram ahead of the June 2026 NEET re-examination, saying the platform had been used by organised cheating operations. The restriction ran until June 22 and affected a service with roughly 150 million users in the country. (Reuters)

Most of those users had nothing to do with exam fraud. They were trying to reach work groups, family chats, customer channels and classes like mine.

The response was immediate. On June 17, downloads of India’s leading VPN apps reached about 919,000 in a single day, the country’s largest VPN-download surge of 2026. (The Times of India)

That rush explained the question appearing everywhere:

Abhi India mein kaunsa VPN sach mein chal raha hai?

The temporary block has ended, but the question remains useful. When access changes without warning, people do not need a theoretical winner or the provider with the longest feature list. They need a connection that works on the network in front of them.

The famous VPN connected, but Telegram did not

I began with the large provider already installed on my laptop.

It was the sensible first choice. The company had years of public history, a large support operation and servers across dozens of countries. I had used it successfully while travelling, and its speed tests at home were consistently strong.

The app connected in less than five seconds.

Telegram continued showing Connecting…

I selected Singapore because it was relatively close. The group list appeared, but the messages did not. I changed to Germany. Telegram opened long enough for me to see twenty-six unread questions, then stopped before I could post the class link.

Next, I tried the provider’s fastest-server option.

The speed test showed 176 Mbps.

Telegram still would not send one message.

For a moment, I wondered whether Telegram itself was overloaded. Then I tested the same setup on my phone. One VPN route worked briefly on coworking Wi-Fi, but failed again when I moved to Jio.

The provider had plenty of servers. What it had not given me was one route I could trust across the two Indian networks available to me.

That distinction matters because internet filtering in India is not always identical across providers. Research comparing major Indian networks found that ISPs used different blocking methods and did not consistently restrict the same destinations in the same way. (Singh)

The everyday version is simpler: an app may open on one network, fail on another and behave differently again on home broadband. A brief public discussion about Telegram captured exactly that mismatch across Jio and Airtel connections. (Reddit)

That was enough to change my approach. A VPN could work perfectly for someone on another ISP and still be useless to the person reading their recommendation.

I did not need another testimonial.

I needed to start the class.

Changing the country did not change the tunnel

I tested three more locations.

One opened Telegram Web but not the desktop app. Another loaded old messages but would not upload the 38 MB exercise file. A third stayed connected for two minutes and then returned to the same grey status bar.

Each server change felt like progress because the country name changed. In reality, I was sending the same familiar type of VPN traffic through a different exit.

Common VPN protocols can be recognised by their traffic patterns. Research presented at USENIX showed that OpenVPN connections could be identified through packet behaviour and server responses. (USENIX Security) The practical implication was clear: changing countries did not necessarily change what the Indian network saw first.

The large provider’s broad server network was useful under ordinary conditions. During a sudden restriction, it gave me more versions of the same experiment.

At 5:57, a student sent an ordinary SMS:

“Sir, is class cancelled?”

I replied that it was starting late.

I had three minutes to make that true.

The free option opened a page, not the class

A free browser extension appeared near the top of the store results. It promised instant access and required no payment, which was compelling with the clock moving.

I installed it.

Telegram Web opened.

For ten seconds, I thought the problem was solved. Then I tried to upload the class file. The transfer reached 4 percent and stopped. The desktop app remained blocked because the extension covered only the browser, while the video-meeting application continued using the normal connection.

The free route had opened one window while leaving the rest of the lesson outside it.

Its first screen also included advertising consent and an upgrade prompt. I did not have time to investigate its operator, permissions or funding model while students waited.

I removed it.

The failed shortcut clarified the task. I did not need a browser trick or a VPN that could merely display a connected icon. I needed Telegram, the file upload and the meeting application to work together.

The setup also had to survive the moment I left coworking Wi-Fi and depended entirely on mobile data.

The message sent before I checked the server

I opened OnlydogVPN^↗, a smaller app I had installed as a backup.

It did not begin with a country map. The first useful choice was the situation, so I selected the preset for a restricted network and returned to Telegram.

The unread messages appeared.

I pasted the meeting link into the group and pressed Send.

The message moved from one tick to two.

Students began joining the classroom.

I uploaded the exercise file next. It passed 4 percent, then 20, then 60. While the transfer continued, I opened the meeting application and admitted the first students from the waiting room.

At 6:02, the lesson started.

The file finished uploading during my introduction.

That was the first result that answered the search properly. The smaller app had not merely connected to a server. It had completed the task that made me search for a VPN in the first place.

Its restrictive-network preset uses traffic obfuscation, making the connection less recognisable than a conventional VPN tunnel. It also removed the country-and-protocol guessing exercise at the exact moment when I had no time for one.

I could not observe the filtering rules operating inside each ISP. I could observe the outcome across the networks available to me: the conventional routes repeatedly stalled, while the obfuscated route opened Telegram, sent the link and uploaded the file.

Once the class was underway, I stopped looking at the VPN screen.

That was the point.

The real test came when the Wi-Fi failed

Twenty minutes into the lesson, the coworking space lost internet access.

The lights remained on, but the meeting froze and the Wi-Fi icon disappeared. My laptop switched automatically to the phone hotspot.

I expected the class to disconnect.

The students’ video tiles paused. The audio broke for a moment, then returned. Telegram remained online, and the meeting recovered without forcing me to reconnect the VPN manually.

The smaller service uses an HTTP/3-based connection that handles network changes efficiently. (IETF) In practical terms, the tunnel followed the laptop from coworking Wi-Fi to Jio instead of treating the switch as an entirely new session.

That recovery became the second reason the app felt suited to the way people actually use the internet in India. “Working” could not mean succeeding only on one fibre connection while sitting still. It had to include the ordinary movement between office Wi-Fi, home broadband and mobile data.

The class continued.

Afterward, I needed to answer questions from my phone while packing the laptop. The app displayed a verification code that connected the second device without another email-and-password login. Telegram opened with the student messages already waiting.

That smaller convenience appeared only after the main task had succeeded. It simply prevented a working setup from becoming another account-recovery exercise on a second screen.

The service has fewer server locations, fewer independent ratings and a shorter public history than the established provider I tried first. That is its clearest limitation.

But the question that evening was not which company had accumulated the largest network or the longest review archive.

The established provider connected quickly and offered many countries, yet I spent the final minutes before class cycling through routes that could not complete the task. The free extension opened one browser page but failed to restore the full workflow. The smaller service gave me one obfuscated connection that worked across Telegram, the upload and the live meeting—and kept working when the underlying network changed.

That is my answer to which VPN is actually working in India right now.

It is not the app that displays the most flags or happens to work for a stranger on another ISP. It is the one that crosses the restriction on the network in front of you and remains connected when India’s next network takes over.