Exactly. At late nights when I am unsure about whether to sleep or continue working, I simply go to the site and calculate when to sleep. It has become a habit!
I'm completely amazed that people do this to figure out when to sleep. My intent is not to sound rude; I'm a big believer of "do what works" in my personal life and at work. For me it's always so simple: if the following day is a work day, I need to make sure I get at least 7 hours of sleep so that determines how late I can stay up. If it's the weekend, I go to sleep when my body says it needs a break.
I think an app like Sleep Cycle is better than this because it can feel your movement, whereas this is just based on the "average" person's sleep cycles...
I think if you keep it on the outside of your bed (away from the partner) and your bed isn't super responsive, it should work. There are others that integrate with wearables and works in that situation, but the concept of it using motion and not just random time is what makes it work more effectively.
The two appear to get different results. To get up at 7:00 AM, sleepyti.me says to go to sleep at 10:00 PM, whereas jollynap says to go to bed at 10:45 PM. Obviously the latter is taking 15 minutes into account for falling asleep, but it's still an hour difference.
SleepyTime mentions, "The average adult human takes fourteen minutes to fall asleep..." so I compensate with 15 minutes in the calculation. There shouldn't be any major variance. Perhaps an odd calculation with Daylight Savings?
The time is based on your browser's Date function which should match your system clock. I think you might be experiencing a strange calculation with Daylight Savings.