It is a travel router, and what it does is more interesting than the reason it gets sold.
Here is the whole of it. A travel router is two radios wearing one name. Upstream it is a client, and it associates to the network you do not control exactly the way any other client device would. Downstream it is an access point running your own SSID, your own DHCP, your own subnet. In between sits NAT, Network Address Translation, which rewrites every address on your side so it all leaves as one address on theirs.
That single fact explains everything it does well and everything it costs you.
What it does well is absorb somebody else's bad network on your behalf. A captive portal that re-challenges. An association that drops and re-authenticates. A gate that counts devices. Roaming that was never designed into the deployment. One device deals with all of that, once. Everything else you are carrying associates to a stable access point in the same room and never sees any of it. They are not on the hostile network at all. They are on yours.
What it costs you is real and usually gets skipped. Everything now shares one upstream association and one radio's airtime, and you have added a hop and a contention domain of your own making. On a congested network you have not made anything faster. You have made it predictable. Those are different results, and predictable is usually the one you wanted.
Worth checking before you buy one. Plenty of travel routers carry a single radio and repeat instead, so the same radio alternates between talking to the public network and talking to you. It works, and it costs you throughput. I carry the two radio kind.
Then there is the part I got to learn the hard way this week.
I was trying to measure roaming behavior on a network I do not control. One device walking a corridor, a second sitting still as a control. I turned the travel router off before I started, and not for airtime. My devices know that SSID and prefer it, and the control device was sitting in the same room as the travel router. It would have held a rock steady signal for the whole session while never once touching the network I was trying to measure.
A known SSID in the same room beats an unknown one down the hall every time. That is not a flaw in the client. That is the client doing exactly what you would want it to do on any other day.
Which is the part worth keeping. A travel router does not make a bad network good. It changes which device has to cope with the bad network, and it changes what everything else can see. Both of those are design decisions, and both of them will surprise you eventually.
What is the piece of kit in your bag that turned out to do something different from why you bought it?