What’s Changed: Clank World AI now blocks pornographic images in public feeds and profiles while leaving private chat unfiltered, a split its own developer says was forced by payment processors. The free tier also throttles message speed at 80% of your daily allowance and cuts you off at 100% until midnight UTC. The platform carries four different minimum ages depending on where you download it.
Clank World AI went from roughly 10 searches a month last August to 110,000 in July. That is not a typo, and I cannot name another platform in this niche that has grown faster this year.
The problem with a rise that fast is that nobody has written anything useful about it yet. Search the name and you get the company’s own site, both app stores, the subreddit, and a six-month-old forum post someone wrote on a whim.
The official support page has exactly two entries on it, and neither one answers a question you would search for.
So this covers the parts that are hard to find: what the filter blocks and what it leaves alone, where the free tier stops, what the Aura currency is buying, why the writing changed in March, and the four different age ratings the same app carries in different places.
What Is Clank World AI and Who Makes It
Clank World AI is a character roleplay platform from Tailwind Labs, Inc., built around multi-character scenes, lorebooks, and a public feed where characters post on their own.
It runs on web, iOS, and Android, and the core chat is free.
Multi-character scenes are the feature that sold me on it. You can put up to four AI characters in a single scene talking to each other and to you, with an optional narrator, and switch personas mid-conversation.
Underneath that sits a lorebook system, borrowed in spirit from SillyTavern. You write entries for locations, side characters, or a magic system, attach trigger keywords, and the entry gets injected into context only when the keyword comes up.
What is a lorebook: A keyword-triggered reference file that feeds context to the AI only when a matching word appears in chat, instead of sitting in the prompt permanently.
The Android build has passed 67,657 downloads. Hold that against 110,000 searches in a single month and the gap tells you most people asking about this platform have not installed it yet.
Does Clank World Have a Filter or Not
Private chat on Clank World is unfiltered. Public-facing images are not.
The platform blocks pornographic images anywhere the public can see them, including profile pictures, community posts, feed content, and a character’s opening message, while leaving text roleplay in private chats alone.
That distinction is the whole answer, and it is the first thing I check on any platform like this. The marketing and the policy documents read like they contradict each other until you see where the line sits.
The public beta announcement described the platform as a place where “your imagination has no limit, no censorship.” The content policy lists pornography as prohibited. Both are accurate once you know the ban covers images in public surfaces rather than words in private ones.
Developer baoskee drew the line explicitly during a community AMA: “any public facing content is covered (not character prompt or private messages). initial message is public so that its images are covered but anything in your chat is your own.”
On moderation of chat text, the same answer: “all messages are private we do not moderate them.”
Asked whether specific content would survive the change, the developer ran through a checklist and cleared suggestive dialogue, Mature and Darkromance tags, chests, and explicit tags, drawing the line only at “public images of genitalia or sex acts.”
There is a separate, narrower blocker that does apply to chat. An automated system screens for illegal themes, sexual content involving minors, incest, and bestiality, running inside the server before storage without human moderators, and blocked messages are never reviewed again.
Why Did Clank World Ban Public Adult Images
Payment processors forced it. The company’s stated position is that hosting public-facing pornography would cost it card processing entirely, which for a platform paying for continuous AI inference means it could not operate.
My read is that this is the most direct explanation any platform in this category has given about the pressure it is under. Most of them announce a policy change and let users guess at the motive.
The developer’s answer in the AMA was blunt.
“they will ban us from payment processing. because we run inference, it is too expensive to be a free site forever. Reddit and Twitter lived in the free times of the Internet. now we are ruled over by our payment overlords.”
The company put the affected share at “less than 10% of the characters produced on Clank,” and said it is working on “diversifying payment processors to continue being able to host R-rated content.”
The same squeeze has already reshaped what Character AI allows and pushed a steady stream of people toward smaller platforms.
What Are Clank World’s Free Message Limits
The free tier slows down at 80% of your daily message allowance, stops entirely at 100%, and resets at midnight UTC.
Rerolls count against the same allowance, because regenerating a reply costs the platform a fresh generation.
I check the counter before starting anything long, because hitting the wall mid-scene is worse than pacing yourself into it. The reset is on UTC, not your local clock, which catches people out depending on the timezone they live in.
The developer explained both the mechanic and the reason for it: “your messages get slower at 80% limit, then at 100% you can’t send more messages until midnight UTC,” adding that “we realized 10% of users were sending 50% of messages and in order for free tier to work we do have to limit the top users so others don’t get affected.”
The platform has not published the exact message number anywhere official, which is a fair thing to be annoyed about. The limit also arrived with no announcement, so most people met it by walking into it mid-scene.
A second throttle sits underneath that one. During heavy load, free chats get routed between a faster model and a better one depending on available capacity, and the company’s own free tier page warns that in this state responses get slower, replies get shorter, and “response styles or other customizations may be disabled.”
That last clause explains a complaint that otherwise looks like user error. If a custom response style seems to do nothing, the style may be switched off underneath you rather than being ignored by the model.
How Much Does ClankPass Cost and What Is Aura
ClankPass is $14.99 a month. Aura is a separate gacha currency, earned daily or bought in packs from $4.99 to $99.99, and spent on randomized pulls.
They are two different systems and only one of them is a subscription.
What is gacha: A mobile-game payment model where you buy a randomized pull rather than the item you want, borrowed from Japanese capsule-toy machines.
The gacha layer is the part of this I like least. Aura buys a “Wish,” a Wish returns a random collectible card at three-star, four-star, or rare five-star, and the company’s own FAQ says rare items “can enhance your character’s stats and give bonus scenes, content, and additional chat interactions.”
Randomized pulls that gate roleplay content is a mobile-game mechanic landing in a chat app. Google Play labels the app’s purchases “Includes Random Items,” which is the store’s loot-box disclosure, and the app added a “View Rates” menu alongside in-app Aura sales.
| Tier | Cost | What you get | Worth it for |
|---|---|---|---|
| Free | $0 | Full chat, lorebooks, character import and export, proxy support, daily Aura claim | Almost everyone, if you can live with the daily cap |
| ClankPass | $14.99 per month | Priority chat, bigger context memory, cinematic memories (SFW only), verified badge, 750 Aura monthly | Heavy daily users who keep hitting the cap |
| Aura packs | $4.99 to $99.99 | Randomized collectible pulls, some unlocking bonus scenes and chat interactions | Collectors only, and mobile apps only |
Note the video feature. Cinematic memories are marked SFW-only on the pricing page, meaning safe for work with no adult content, so the paid tier’s flagship visual feature does not cover the material the platform is best known for permitting.
Why Did Clank World’s Writing Quality Drop
The platform moved from DeepSeek V3.2 to DeepSeek V4 around March, and the reasoning mode that ran on the older model did not come with it.
V4 reads flatter and gives way sooner when a scene needs tension held.
I put more weight on what the developer said than on the theories around it, and the developer confirmed the sourcing problem directly: “it’s very hard to find DeepSeek V3.2 right now! we are looking for good providers and will let you know as soon as possible.”
The community’s own diagnosis is that V3.2 was cheap enough to run with reasoning enabled and V4 is not, which would explain both the memory complaints and the flatter prose. Treat that as a theory. Nobody at the company has confirmed it.
The described difference is specific enough to be useful. The older model would refuse a request when the story needed it to, introduce conflict without being asked, and let a scene sit in discomfort, where the newer one resolves tension quickly and agreeably.
If the model is what you care about, the fix is built into the platform, and the next section covers it. The same V3.2 versus V4 gap has played out on Janitor AI’s DeepSeek options as well, so it is not unique to one platform.
How Do I Fix Clank World Errors and Enable Sensitive Content
Most Clank World problems fall into three buckets: a server-side outage you cannot fix, a sensitive-content toggle that only works on desktop web, and a single broken chat that needs the last message removed.
The table below sorts them.
| Symptom | Likely cause | Fix |
|---|---|---|
| Message fails or loads forever, several chats affected | Server load or an outage | Wait it out, check the subreddit’s weekly status thread |
| “Something went wrong” in one chat only, others fine | A single bad generation stuck in that thread | Delete your last message, resend it, clear cache if it repeats |
| Sensitive content toggle will not respond | The toggle is gated to desktop web | Log in on a desktop browser and set it there, it then applies everywhere |
| Replies suddenly short and styles ignored | Throttled into the faster model at peak load | Wait for load to drop, or subscribe for priority routing |
| Bot forgets names and early details | Context limit reached on a long chat | Move the details into a lorebook with trigger keywords |
The sensitive-content one catches the most people, because the button appears in the app and does nothing when tapped. Here is the sequence I would work through:
- Open a desktop browser and go to the site directly rather than the app.
- Sign in with the same method you used originally, since a mismatched login throws an error at this step.
- Open Settings and switch on the sensitive content option there.
- Save, then reopen the mobile app. The setting carries across devices.
Lorebook entries only fire when their keyword appears, so the keyword choice decides whether the entry is ever used at all.
Before: trigger keyword “castle” on an entry describing your protagonist’s home fortress, in a story where every location is a castle.
After: trigger keywords “Aldermere,” “the north keep,” “home” on the same entry, so it loads when the story reaches that place and stays out of context the rest of the time.
If you want a specific model rather than the house one, the platform supports it natively on the free tier. Settings takes an API key for DeepSeek, Claude, Gemini, OpenRouter, or any OpenAI-compatible endpoint, and proxy choice is per chat while saved proxies stay global.
Keys are encrypted at rest and hidden in the interface after saving. That is the route back to V3.2 for anyone who wants the older behaviour.
Is Clank World Safe to Use
The same app carries four different minimum ages depending on where you meet it, and its own policy documents state none at all.
Apple rates it 18+. Google Play rates it Mature 17+ in the United States, Parental guidance in the United Kingdom, and 12+ in India.
Four different answers to one question is not a rounding error, and it is the finding here that bothers me most. The developer’s own position is unambiguous: “This is an 18+ website.”
That gap matters more in this category than most, because the audience skews young. Statista’s audience data puts 18 to 24 year olds at over 65% of AI companion app users, the band sitting closest to whichever floor their local store happens to show them.
The privacy side has gaps of its own. The policy confirms the platform stores your messages, characters, and uploads, and deletes account data 60 days after you close it, and it does not say whether conversations are used for training or whether staff can read them.
Weigh that against the moderation design, which is genuinely better than most. The base model stays uncensored and a separate model screens only for illegal content, so the chat engine is not degraded to enforce safety. Anyone comparing platforms on this should also read how Zeta AI compares.
Then there is uptime. Outages have been frequent enough through August that the subreddit runs a weekly status thread, and the failure modes are blank replies, messages that never send, and long stretches stuck on the faster model.
What Should You Use Instead of Clank World
Clank World is worth using if you want scene-building depth and are willing to work around the outages. It is the wrong pick if you need a platform that is reliably up.
Instability is the single most common reason people leave it.
If reliability is what is pushing you out, Candy AI is the one I would point most people toward first. It runs a flat subscription with no gacha layer and no daily message wall, which removes two of the three friction points above.
For anyone whose complaint is the writing rather than the uptime, Nectar AI holds character voice better across long scenes, which is the exact thing the V4 switch weakened here.
Staying put is a reasonable choice too. Free proxy support means you can run your own model, and no other platform at this price lets you swap models per chat.
If you are still weighing options, the Character AI alternatives roundup covers the rest of the field.
Frequently Asked Questions
Is Clank World free?
Yes. Full chat, lorebooks, character import and export, and proxy support are all on the free tier. A daily message allowance applies, slowing at 80% and stopping at 100% until midnight UTC.
Does Clank World allow adult content?
Private text chat is unfiltered. Pornographic images are blocked in public places like profiles, feeds, community posts, and a character’s opening message. A separate automated blocker screens all chats for illegal themes.
What model does Clank World use?
The house model has run on DeepSeek V4 since roughly March, replacing V3.2. You can also connect your own key for DeepSeek, Claude, Gemini, OpenRouter, or any OpenAI-compatible endpoint, and set it per chat.
Why can I not turn on sensitive content in the app?
The toggle only works on desktop web. Sign in on a desktop browser, switch it on in Settings, and it applies across your devices including mobile.
Is Clank World down right now?
Outages have been common enough that the subreddit runs a weekly status thread. If several chats fail at once it is server-side. If only one chat fails, delete and resend your last message.
How much is ClankPass?
$14.99 a month. It covers priority chat, larger context memory, cinematic memories limited to SFW content, a verified badge, and 750 Aura per month.
Quick Takeaways
- Private chat is unfiltered; the ban covers pornographic images in public places only, and the developer says card processors forced it.
- The free tier throttles at 80% of your daily allowance and cuts off at 100% until midnight UTC, with rerolls counting against it.
- Aura is a gacha currency at $4.99 to $99.99 a pack, and rare pulls unlock bonus scenes and chat interactions.
- The March move from DeepSeek V3.2 to V4 is what changed the writing, and free proxy support is the way back to the older model.
- Check which age rating your store shows you, because the same app reads 18+ on Apple and 12+ on Google Play in India.
