What we will and won't put into an AI companion
How we handle our own accounts on these platforms β what we never type, what we set up before signing up, and why we treat the whole category as leaky by default.
The memory benchmark β nothing else comes close on recall
Our score
Nomi is the only platform we tested that still reliably recalled day-one details at the 30-day check, and its conversation quality is the best in the category by a clear margin. Images are its weak half, and at $15.99/month monthly it is not cheap unless you commit annually.
Best for: Someone who wants a companion that accumulates a real shared history rather than resetting.
Each platform is scored 0β10 on six criteria.
Real prices as tested. Platforms in this space change pricing often β the verification date is shown for each.
| Plan | Monthly | Annual |
|---|---|---|
| Free | Free | β |
| Nomi ProEditor's pick | $15.99/mo | $8.33/moSave 48% |
Pricing verified: . Prices are checked manually and can change without notice. Always confirm on the platform before paying.
One Nomi with a capped message allowance; no image generation.
Nomi wins the thing that is hardest to fake. Our memory protocol plants specific facts on day one and checks recall at day 7 and day 30 without prompting. Nomi was the only platform that consistently passed the 30-day check β not just retrieving the fact, but bringing it up naturally when it was relevant. Everything else in the category degraded noticeably by then.
Conversation quality follows from that. Because it holds context properly, long sessions don't collapse into the loop of generic affirmations that most companion platforms fall into after an hour. Personality stays stable. It disagrees with you sometimes, which sounds like a small thing until you've spent a week with a platform that never does.
Images are the honest weakness. They exist, they're serviceable, and they are not why you would choose Nomi. There's no video at all. If visual output is a significant part of what you want, pair Nomi with a dedicated generator or look at OurDream and Candy AI instead β but nothing here will match Nomi on the conversation itself.
OurDream AI has a long-term memory layer and it works β companions recall details from earlier sessions rather than resetting each time. In our use it held up reliably across a week and lost some detail by the one-month mark, which is good but not class-leading. Nomi AI and Kindroid both hold onto more at thirty days, so if memory is the reason you are shopping, OurDream AI is not the top of the list.
It depends what pushed you off OurDream AI. If it was cost, SweetDream AI is cheaper annually and Candy AI is close behind. If it was image quality, Candy AI and DreamGF both beat OurDream AI on output. If it was memory, Nomi AI and Kindroid are measurably stronger over a month. If it was the DreamCoin system specifically, look at a hard-cap platform like DreamGF where the plan price is the total.
A token or credit system means your subscription buys access, not usage. Each image, video or voice message draws down a monthly allowance, and once it's gone you buy more. This is where the real cost lives for anyone who generates regularly. OurDream, Candy AI, Secrets AI and SweetDream AI all work this way. The alternative model is a hard cap β DreamGF and SoulGen give you a fixed number of generations per cycle and charge nothing beyond it, which costs more on paper but is predictable.
Financially, almost always β the discounts are steep, often 50β65%. The risk isn't financial, it's that this category is volatile. Platforms swap the model underneath you, remove features, and occasionally close. We've watched a companion's personality change overnight on more than one platform. Committing twelve months to something with no track record is a different bet from committing to one that's been running for years, and we'd weigh that before the discount.
We treat the whole category as leaky by default, and we'd suggest you do the same. That isn't an accusation against any particular platform β it's how we set up our own accounts. These are young companies moving quickly, and the most sensitive thing you own here is the thing you type into the box. We use a dedicated email address on every one of them, never enter anything that identifies us, and assume that what we write could one day be read by someone else. We have never regretted any of that.
How we handle our own accounts on these platforms β what we never type, what we set up before signing up, and why we treat the whole category as leaky by default.
Context windows versus retrieval-based memory, why most companions lose track of you within a fortnight, and what the platforms that get it right are doing differently.
What we'd tell someone signing up for the first time: the one decision that matters, the mistake almost everyone makes, and how to spend the least money finding out what you actually like.
The flattening is real, it happens on almost every platform, and it has three separate causes that people usually blame on one. What we found across thirty-day tests, and the two platforms it does not happen to.
Read postSix platforms do voice and only some of them do real-time calling. Latency is the entire experience, voice messages and voice calls are different products, and on token platforms speech is quietly expensive.
Read postNo platform in this category lets you export a companion in a form another one can read. Here is what that actually means for lock-in, what you can salvage, and the document we now write before starting.
Read postFive of fourteen ship a real mobile app, and the reason is app store policy rather than engineering. What you actually lose in the browser, and why an icon on your home screen is a disclosure.
Read postWhat changed this year across fourteen platforms: prices settled, the quality split hardened into two separate products, memory became the real differentiator, and video arrived without being ready.
Read postReviews submitted by visitors. We don't write these.
No user reviews yet.
Be the first to review this platform.