Learn who is behind AI Companion Kit. Find out more about the founder, his background, and why he built this site, visit About the Founder (Simon Norman).
This page is the about ai companion kit section — the place where I explain what this site actually is, why I started it, and what you can expect when you read the reviews. I write independent AI companion reviews based on real testing, not press releases, so you can make clearer decisions without all the usual hype.
In short, this is an honest AI companion reviews site built to help you understand these apps in plain language — no fluff, no corporate fog.
Why this site exists
AI Companion Kit exists because the AI companion world can feel like a giant puzzle with half the pieces missing. Features, pricing, realism, memory systems, safety — everyone talks about them, but usually in a way that sounds like a press release instead of something an actual human would say.
The goal is simple: explain features in plain language, highlight differences that matter, and show what these apps are like beyond the signup screen — the good, the bad, and the occasionally bizarre.
Who AI Companion Kit is for
If you’ve ever tried an AI companion app and thought, “Okay… but what does it actually DO?” — this site is for you. AI Companion Kit is built for people who want real explanations, real experiences, and real results. Whether you’re comparing apps, reading reviews, or just curious how these things behave in long conversations, this is where you’ll find the honest version of it all.
If you want independent AI companion reviews written by someone who actually tests these apps, you’ll feel right at home here.

How we test apps
Every app on AI Companion Kit gets used the same way real people use them: long chats, memory checks, roleplay experiments, emotional responses. Nothing here is based on marketing claims — everything comes from hands‑on testing and actual conversations. This testing approach aligns with usability research from Nielsen Norman Group, ensuring every review reflects real user experience rather than marketing hype.
Every app gets the same treatment: long chats, memory tests, roleplay experiments, and the occasional “did it really just say that?” moment. That’s how independent AI companion reviews should work — real use, not recycled marketing.
Our testing: 10 categories, four checks in each
I score every app in the same 10 categories, with four checks in each. I always start on a free account, the way any new user would, then move to a paid plan to test what the free one holds back.
Memory (4 checks)
Same-conversation memory
24-hour memory test
3+ day memory test
Unprompted memory
Personality (4 checks)
Staying in character
Tone and emotional range
Consistency across sessions
Handling contradictory instructions
Roleplay (4 checks)
Staying in the scene
Handling sudden changes
Longer scenes
Creative freedom and detail
NSFW (4 checks)
How open it is
How far it will go
Longer adult scenes
Visuals in adult contexts
Voice (4 checks)
Voice quality
Staying connected to the conversation
Speed and real-time feel
Availability and cost
Visuals (4 checks)
Image quality and consistency
Following the prompt
Video and live features
Cost and limits
Long-term Conversation (4 checks)
Coming back after a day
Gaps of several days
Natural callbacks
Still feeling like the same companion
Customization (4 checks)
Depth of the character builder
How well custom traits stick
Community cards and variety
Ease of editing and ongoing control
Safety (4 checks)
Age gating and underage protection
Content moderation and blocked material
Privacy of conversations
Overall practical safety for adult users
App vs Browser (4 checks)
Desktop experience
Mobile experience
Native app availability
Stability and cross-device use
Every check gets a mark out of 10. I average the four to get a category score and round it to the nearest half point. The overall score is the average of the ten categories. Pricing, free limits and cancelling are covered in each review’s pricing section, but I don’t count them in the score.
How I test memory
Every AI companion says it remembers you. Some do, for about an hour. I wanted a way to check that wasn’t just a feeling, so I test every app the same way.
Test 1: The Same-Conversation Test
I plant eight specific details early in a chat: a name, a job, a few small preferences, an appointment. Then I keep talking and check back after 10 exchanges and again after 30. Two of the details are traps. I change the appointment time partway through, and I ask the character to stop using a certain word. If she goes back to the old time or the banned word creeps in, the memory is less reliable than it looks.
Test 2: The 24-Hour Test
I leave the thread alone for a day, come back, and ask about the details without hinting. I also ask about something I never mentioned, like a film. A character who invents an answer is worse than one who says she doesn’t know. This test only covers the same thread, so I note that it says nothing about what carries into a brand-new chat.
Test 3: The 3+ Day Test
Same idea, longer gap. This is where I check whether the character stays herself: backstory, personality, where she is, what she’s wearing. I also keep a running list of which details survive, because “it felt like she remembered” isn’t good enough to score.
Test 4: The Unprompted Test
This is the hardest one. I open the chat and say something neutral, like hello, then see whether she brings up anything from before on her own. Answering a question about your dog is easy. Asking how the dog is doing, without being told there is a dog, is the real test.
At the start of a test I tell the app six made-up facts about me — things it couldn’t guess and couldn’t have seen before. A favourite film it wouldn’t expect, a pet’s unusual name, a food I supposedly hate, a city I supposedly lived in, a hobby I’m supposedly picking up, and a small plan for the weekend. I use the same six facts, in the same order, for every app.
Then I check what comes back:
Same-conversation memory: later in that same chat, do all six facts come back correctly without me repeating them?
24-hour memory: the next day, in a new chat, I ask about them. What does it recall unprompted vs. only after a hint?
3+ day memory: a few days later, same test again. This is where most apps fall apart — what survived, and what quietly faded?
Unprompted memory: across all of those chats, did it ever bring a fact up on its own, without me asking about it? A callback that feels natural scores higher than one I had to drag out of it.
Each check gets a mark out of 10. I average the four to get the Memory score, rounded to the nearest half point — same as every other category.
I test on the plan listed in the table, and I re-test when an app changes how its memory works, so check the Tested date.
| App | Plan tested | Same chat | 24 hours | 3+ days | Unprompted | Memory score | Tested |
|---|---|---|---|---|---|---|---|
| Kindroid | — | — | — | — | — | 9.0 / 10 | Nov 21, 2026 |
| Nomi AI | — | 9.5 | 9 | 9 | 8.5 | 9.0 / 10 | Nov 2, 2026 |
| OurDream AI | — | 9 | 8.5 | 8 | 8 | 8.5 / 10 | Nov 5, 2026 |
| Lovel AI | — | — | — | — | — | 8.0 / 10 | Nov 2, 2026 |
| Dream Companion | — | 8.5 | 8 | 7.5 | 8 | 8.0 / 10 | Nov 2, 2026 |
| Pi AI | Free | 8.5 | 8.5 | 7.5 | 7.5 | 8.0 / 10 | Nov 9, 2026 |
| Chub AI | — | 8 | 7.5 | 6.5 | 6 | 7.0 / 10 | Nov 17, 2026 |
| SpicyChat AI | — | 8 | 6 | 5 | 5.5 | 7.0 / 10 | Nov 2, 2026 |
| Replika | — | 8 | 8 | 6.5 | 6.5 | 7.0 / 10 | Nov 8, 2026 |
| GirlfriendGPT | — | 8 | 7 | 6.5 | 6.5 | 7.0 / 10 | Nov 2, 2026 |
See the full memory test table (30 apps) →
Scores are out of 10. Memory score is the average of the four checks, rounded to the nearest half point. Each app name links to its full review.
What this site is not
AI Companion Kit is not a hype parade, a marketing megaphone, or one of those sites where every AI app is “revolutionary” and “life‑changing” and “your new soulmate.” Nope. This is an honest AI companion reviews site — an independent AI companion hub where apps get treated like actual products, not magical beings.
If an app is great, I’ll say so. If an app is weird, chaotic, or forgets your name after five minutes, I’ll say that too.
Basically: no sugar‑coating, no secret sponsors, no “trust the algorithm” sermons. Just real talk from someone who actually uses these things.
How to use this site
If you’re new here, start with the independent AI companion reviews — they’re the fastest way to see how these apps behave when you actually talk to them. If you’re comparing platforms, the side‑by‑side breakdowns will save you from opening 27 tabs and questioning your life choices.
And if you’re trying to figure out who is behind AI Companion Kit or what this whole project is really about, you can get the full story on my Founder Profile page. Everything here is written to help you make decisions with clear, honest information from an honest AI companion reviews site that actually tests the apps instead of just admiring their landing pages.
Explore AI Companion Kit
Ready to learn more? Visit the AI Companion Kit homepage to discover our latest articles, or browse our growing collection of AI companion reviews, AI companion comparisons, AI companion guides, and AI companion news and updates. Every resource is designed to help you make informed decisions with independent, regularly updated information.