The bill
I went to Iceland in late 2023 for what was supposed to be a week of glaciers and hot springs, came home, and got a $400 phone bill. Not a typo — four hundred dollars.
The breakdown was: roaming charges. I'd kept my US carrier's roaming pass enabled because I'd been told "it's just $10/day." It is $10/day, technically, until the day you accidentally let a Zoom call run for 40 minutes from a hot spring in the highlands. Then it's $10/day plus $1.50/MB of streaming over the daily-pass cap. I learned about that pricing tier through the bill itself.
I'd prepared for the trip. I'd packed thermal layers and traction-spike grips for boots. I'd downloaded offline Google Maps. The one thing I hadn't done — because no one had quite explained how to do it correctly — was buy a local SIM in advance. The kiosks at Keflavík airport were closed at 6 a.m. when I landed. By the time I figured out the alternatives, the daily roaming pass had already kicked in. By the time I got home, $400.
The eSIM moment
A friend told me later, "you should have bought an eSIM. Twenty bucks, takes ten minutes, you'd have been fine."
I tried buying one. The experience was: open a comparison site, get hit with 60 plans, none of which clearly explained which carrier they rode, what coverage looked like in the highlands vs Reykjavík, or whether 3 GB was enough for what I was actually doing. The marketplaces were either too sparse (here are 5 options, take your pick) or too dense (here are 200, good luck).
Both failure modes have the same root cause: nobody on the supply side actually understands what the buyer is trying to do. The buyer doesn't want a "1 GB plan for Iceland." They want "data that works for a one-week trip with hot springs and a couple of remote drives." That's a different question.
What we're trying to fix
The bet behind Roamingly is that the right interface for buying a travel eSIM is not a comparison grid — it's a conversation.
You tell us "Tokyo for nine days, working remotely from a co-working space, Slack and Zoom every morning" and we surface the three plans worth your attention. We explain — in plain English — why one is cheaper, why one has better Hokkaido coverage, why one is unlimited but throttles after 1 GB/day. Then you pick.
The model behind that conversation is a Llama-class LLM running stateless inferences. The catalog under it is 320+ live plans across 29 countries. The result is — I hope — that the buyer never has to ask "wait, which carrier is this?" or "is that fair-use unlimited or real unlimited?" because the AI just tells you.
That's what we couldn't find when I needed it. That's what we built.
What's next
We're a beta. Real customers are buying. The catalog will expand. The next pieces in flight:
- Email confirmations + a real admin tooling layer (so customer support stops being me, on Telegram, at 3 a.m.).
- A proper currency layer: live FX, prices in your local currency, USD anchors for trust.
- Coverage maps that show carrier-specific reach, not just "country covered."
- Razorpay webhook integrity so a closed browser tab never costs anyone their plan.
If you have a trip coming up and want to skip the comparison spreadsheets, the chat is open. And if you've ever come home to a $400 phone bill — I see you. We built this for both of us.
— Kislay