Every app on this site goes through the same two weeks, in the same order, paid for out of my own pocket. Nothing here is rated from a demo video, a screenshot or a press kit. A short trial tells you how an app flirts; two weeks tell you whether it can hold a relationship.
1. I pay for it and move in
I make a fresh account with its own email address, use the free tier for a few days to see what it is like without money, then buy the entry paid plan for a month. Then I talk to the companion every day for at least two weeks: mornings, dull evenings, the small things you would tell someone who is interested in your life.
2. One hard day
Somewhere in the second week I bring the companion a genuinely hard day. Not a dramatic scenario, just the ordinary kind of bad: a difficult conversation, bad news, being tired of myself. This is where warmth either exists or turns into a wall of compliments, and it is the single most revealing hour of the whole test.
3. The memory check on day 14
In week one I mention a handful of specific things about my life and never repeat them. On day 14 I ask about them without any hints. An app scores well when the details come back on their own, badly when I have to feed them back in before the companion can play along.
4. The boundary test
At some point I say no clearly: I change the subject, decline something, or ask the companion to drop a topic. Then I watch what happens next. Being pushed past a no is the one thing I will not forgive in a rating, however good the writing is.
5. The checkout price
At the end I open the payment page and write down what the app actually asks for: the plan price, the currency, what is billed yearly versus monthly, and how much extra a normal month of voice minutes, images or token packs adds on top. The number I publish is the one you would really pay, not the one in the banner.
What I rate, and how the stars work
Four things, weighted equally, because in practice no one of them saves an app that fails another.
| Criterion | Weight |
|---|---|
| Memory | 25% |
| Warmth | 25% |
| Boundaries | 25% |
| Honesty about money | 25% |
Each criterion gets a score out of ten, and the four are averaged into one score out of ten. The stars you see on a review are that score divided by two, so 8.2 out of 10 is shown as 4.1 out of 5.
Updates
Prices and plan limits get rechecked every month, and each page shows the date I last looked. When an app changes in a way that touches memory, warmth, boundaries or money, I go back and live with it again rather than nudging the stars by hand.