itch.io

Redoni i think you're still missing the most definitive proof here. You're saying we can't infer bots just from patterns, but you are forgetting that Ofihombre literally stated the DBolical administration already confirmed the issue

That there is a confirmed bot problem on this site does not mean, you can infer it just from patterns. One does not follow from the other.

And most importantly, I do not happen to see those alleged patterns. What I see is perfectly normal traffic to download ratio for an indie game on a less popular platform.

When the views go up, so go the downloads. That directly contradicts this assumption that the views are mostly patterns of bot traffic.

So from what I know currently, I assume that most of this bot traffic is either removed from the statistic, or the games of Ofi are just not hit by that fake traffic.

I did check out a few other games from the currently popular sorting of that site. They had even less traffic than Ofi's game. And no such patterns.

that Ofihombre literally stated the DBolical administration already confirmed the issue 

Then why not simply cite this. Instead a non conclusive mess of oddidites is presented as evidence. Did those admins say how much such traffic there is, and if they remove such traffic from statistics?

 We aren't arguing about whether the site is completely useless to find a game; we are arguing that the stats system is broken

No, we do not. You might. Ofi is not. And I am pointing out the flaws in Ofi's logic.

The thread's title is:

IndieDB/ModDB is a bot/crawler farm! Don't public any game/mod!

I think that is supposed to mean publish and not public. Ofi calls for other people to not use this platform and claims it is used to simulate traffic, rather than there being genuine traffic.

Ofi is not arguing that the statistics of that site are broken and should not be used for marketing analysis. It is a call to boycott that site.

but rather for those bots and crawlers that impersonate Mac users and are counted as "guests"

And you have any proof for that? What's the motivation? Who is gaining from this? Who does this? Why impersonate mac users? Would it not be more clever for bad bots to impersonate the expected user agents representing a general audience? 

You just said the site was previously more popular. Maybe it is still popular for mac users for whatever reasons. At least more popular than for windows users, that is. A few hundred users is not a big number to do statistics. A single school or university fitted with computers for the students could dominate such a statistic.

What I was trying to demonstrate in the patterns between the statistics of my games, Padman and Garry's Mod, are due to the bots from guest counter

You build a premise that itself is based on assumptions. What you showed as evidence is not conclusive and the big flaw is: even if there was a correlation between games, you could not prove that it would be because of a bot traffic wave hitting those games.

So am I to believe that this alleged bot wave traffic selectively hit some games with 2000 hits and others with 200 hits and still others not significantly at all? How does your hypothesis explain this mismatch of supposed bot activity?

This is almost as bad as: I see lights in the sky, therefore aliens. You just jump to your conclusion, it must be the aliens bots that are visible there.

I am critizising your chain of logic here. Your observations are not explanations. They are observations. And the only thing you show here is, that the view to download ratio is rather bad on that site. And even that is not true, it seems.

AI are good for that sort of summary. I asked what ratio to expect on indie gaming sites. It gave me these numbers.

1–3% downloads per view — common for an average listing

3–10% — solid performance

10%+ — very strong, usually a highly targeted audience, compelling page, or game-jam/viral traffic

Below 1% — not necessarily bad; often caused by broad exposure, weak store-page messaging, paid games, or visitors who already own the game

And your so called unusal spikes of bot activity fit this ratio. You do have like 1 download per 100 views. Sometimes 4 downloads for 53 views. Then one day 216 views and 3 downloads, next day 162 views and 1 download. And the traffic below 50 often has 0 downloads.

Nothing in the data of that Randy game seems suspicous to me at the moment. You do not have 1000 views and 0 downloads. And if your views pass into the hundreds you do have a proportional download ratio near 1%, which is to be expected.

Post your ticket number here. It might speed up the support issue.

And of course your account is still connected to your payment options. That's why practically all payment options have additional security features at the moment of paying. So that people knowing your card numbers cannot just simply use it. You should periodically check your credit card for suspicous activity anyway. And also make sure you do not use the same password on other sites that were associated with the email adress of your stolen account.

If the game is no longer available on Itch, recovering your account will not give you that game. Unless it is one of the cases where the developer made the page unavailable for new buyers but it is still available for previous purchases.

Read that again. 

whenever a developer makes a minor change to the page itself like editing the text description, updating tags, adding new screenshots

The update date of a page is the update date of a page.

Is this about the Itch app? That would be front end. Propagating or even downloading phantom updates, while nothing was there to be downloaded, seem like a bug.

Complaining that the statistical data on that platform is useless is not the same as calling out the platform as useless. When searching a game from OP that game's indiedb page only appeared at #9 in results. That "low" number would be the indicator that indiedb is far less important than Itch or Steam. But it does show up. People preferring that site can go there and look up information about the game there. If they download there, is another question. I would not. I would go to other places for the actual download.

My main issue was the chain of logic. It does not hold up. You can't infer bots from those patterns and the other oddities. The only thing you can direcly see, is, that people do not use indiedb to download a game very much and there are a lot of other plausible reasons for that. Also, there actually was a rise in downloads with the November spike correlating with the bigger spike in November. 

Now there, arguing that views on that site are useless and maybe suspect, if they do not result in downloads and comments, is something different than arguing that you see "patterns" and because of patterns those views are bots.

You do not need to construct a pattern based bot invasion. Because those patterns can be explained by other means. Even if you find games that show similar patterns, that does not prove anything specific. For video games there are seasonal effects. And even for a small platform, a popular game can attract new people to the platform itself and those new people might check out a lot of games at the same time, resulting in a lot of spikes for a lot of games at the same time.

There is two things here, and you do not separate them. One is bots visiting the site. For that you need just cite the relevant information from the site admins talking about that. Or just look at the bot counter they show. I do not know what these bots are for, and why it is 600+. And why the site knows they are bots, but does not block them. The only question here is, if those 600+ bot crawlers are shown in the statistic as views. I do not think so. It is 600+ and the spikes you show are well below 600. Also, it would be very unusual to count a crawler as a visitor.

The other is, that views are not resulting in downloads. The problem is, the first will not prove the second. Unless you can convince people that all the traffic is from bots. The number of views from bots is bound to be an absolute number and not a percentage. There is a countable number of bots, even if you do not know the number. The kind of bot that does illegitimate things and does not identify itself as a bot. Crawlers visiting a site is normal and would not show up as a visitor, unless it is malconfigured or intentional.

The gary mod statistic has a baseline of about 500. The height of the spike is 2500. So how many bots are there? And why is your spike not 2500, but 200? Why do the bots selectively target games and ignore others? Even if those crawlers are shown in the view statistic, maybe this just reflects an increased interest in the topics of a game. Those crawlers do summary snippets and other things. If people are interested in garys mod, more crawlers will try to summarize the pages of garys mod.

So let's talk about "downloads". The statistic you showed does have increase in downloads at those peaks. Up from 0 to a few and for the November spike even rising with the spike. Is there an expected conversion rate for that site? Do people go to this site to actually download a game? When I go to an example indie game like this one, I see a "download" button that does not show files, unless expanded. And it shows a link to Steam and to Itch.

This is my personal preference of course, but that db site is ad-ridden and almost unuseable. It ranks quite low in my own trust, and frankly I do not know that site and do not know how to navigate it. That game in particular has 8 different demo versions to download. There is not even version numbers. Just a dump of downloads to sort through. And I did mention, that the download needs to be expanded, while there is an attractive Steam button below? ;-)

 Would I download from there? Nah. I would check the game out on Steam. Or maybe Itch. Also, the game is in early access. While I am not a user of that indiedb site, I probably would not go there to download games that are still in alpha or beta. Also the site has DB in it's name. I do not go to imDB to watch movies, I go there for information about the movie.

So is this free advertisement by having your game on indiedb a waste of time? Crawlers can find it there. And according to your arguments, they do that a lot and inflate your view counter. You are just disappointed because you do not see a direct increase in downloads. What do you think those bots do with the information? Do you think they increase the view counter just for fun? That information goes somewhere, and can result in someone knowing about the game and maybe even checking it out - at whatever place.

It will also appear in search results. This currently gives me these answers: Steam. Itch. SteamDB. Weebly Homepage. Youtube. ModDB. Nekoweb. Reddit. indiedb. Gamespress.

I think that because you didn't quite understand my point

I literally told you so...

and my games weren't the only ones affected

I did pull up some games from that page and had a look. I did look at the previous year. I did not see what you seem to see.

There you will also notice the same pattern that I had detected in my stats.

I do not. The spike in July is missing.

Your argument, as far as I understand it, is this:

You see spikes of traffic on that site. You see things that seem strange. You assume that this is mostly bot activity. You know that the site owners do something against bot traffic. It is not good enough in your eyes. Therefore you advise people to not use that site.

Thing is, you do not need to construct a premise of bot activity. You can just cite the site admins where they do adress this and claim they are aware of bot activity and try to curb them.

But then you go on to call for an action. That people avoid that platform, because the bot activity is not curbed enough. What the? Why?

I am not sure what your point is and what you are a showing here.

People using private browsing features will not show a referer.

This site has embed buttons for promotions. That strange referal looks like a faulty embedded one.

You are on Steam. Those spikes seem to correlate a bit with the seasonal sales Steam has. Even without that, seasonal ups and down are to be expected and you have a portfolio of games and not ony one game.

Whatever statistic you are showing here, this is max 200 daily views. The average daily views is more like 25. This is not a stable regular counter. It will be highly susceptible to any kind of promotion and random fluctuation. This is not a news site people check daily. If they visit your projects, they visited it and need not visit for some time.

And if you compare to other projects, do you find those same patterns? I did not. If this were mostly bot traffic, why target your pages? Why not everyone?

The usernames you posted (do not do such things, btw.) are users that are logged in. The page has a guest counter. And bot counter.

Statistics

Online 8

Guests 2,006

Bots 630

So your illegitimate bots would not be members, but guests or bots. No clue what the regular bots do. And why so many. 

The only strange thing here is the popularity of mac os. They only have a market share of like 25%, and Windows 60%. Are the games posted on those sites maybe more popular for mac os users? Other users going straight to Steam or even Itch?

So unless this site is charging you money for "visitors", I do not see the problem here. Your "evidence" does not convince me of your traffic being mostly bots. And my strongest critique is, that you only looked at your own traffic. Look at other projects. I looked at some that pop up in the popular today bracket and checked their statistics. Some of those had a max peak of 20 for the last year. And I did not notice any strange patterns, like from 20 to 200 several times. Some had singular "spikes", but going from 10 to 100 is only a spike, because the numbers are small. Statistics on small numbers are fickle.

Everyone talks about ai, ai assets, ai disclosure. But what about ai agents?

Is using a bot on Itch allowed? I would not think so, but I did not find this explicitly stated in the tos.

But the tos is written from the perspective that the publisher is a human. It attributes human qualities to the publisher, like the ability to enter an agreement. I am not a legal expert and this is tos anyway, but I do not think you can delegate certain actions to a tool. Not when interacting with a public audience, like publishing projects or posting comments.

It is one thing to use ai for assets on your hobby project or use a coding assistant, but another thing to use an ai agent to post hundreds of projects a month, trying to sell those and even create comments, faking to be a communicative publisher.

Sure, you could mark the games as ai made, but tagging those games as having ai assets does not really capture that the full game was actually ai assembled, the page design ai made, and the game even ai published.

Anyway, I do not think this is something Itch would want to have on the platform. And I consider using a bot to act on your behalf and simulating human interaction with the platform to be an abuse of Itch's services. It is disrespectful and foremost: it is spam. It clogs up Itch capacities and distorts the catalogue. Users assume the things they see are from indie publishers here. Not from spamming ai agents.

I think Itch needs to adress this firmly, clearly and early. Else some few ai agent users will abuse the situation and will thoroughly flood the browse pages. The many projects by ai enthusiasts is bad enough, but the output of ai agents doing this is another 100x on top of that.

How to protect against those spammers? Sure, you could try using the no-ai tags. But this will also sort out all the regular people using some ai - and for code that is industry standard in the professional world. Those ai agent users are not hobby developers. Their activity is not creating games, it is supervising their ai agents.

There is no propagating through the network. Just because you uploade a patch does not mean anyone downloaded it.

Also, maybe you did not actually fix the bugs for general usage, but only for special circumstance - or you just messed up the build.

/search?q=CnrkiiART

Your things are not indexed (yet). And at least one is in quarantine.

/t/4120453/unofficial-search-and-indexing-faq

Is this a local system or an interface to other services?

Because you will not create an "ai" from scratch as a project. So are you adapting free model for local usage or is this yet another client interface. Or even one of those ai agents?

Slop essentially means low effort. Outside of software it can mean certain types of quickly made meals. It can also mean drainage water and waste.

If a dev releases 30 projects a month with ai, that is mass produced and ai slop. And you will find such "developers" that release even more than that. Though I have seen people do that without ai.

I have no issues with a hobby dev using ai to help on their hobby project. But there are accounts that use ai to churn out several projects a day and even with a price tag. This is not what Itch was made for. This garbage is not only ai slop, it is spam that clogs up Itch services.

One of those ai slop creators even has the audacity to post ai response to topics here in forum. Itch needs to ban this kind of behaviour. And ban it hard. With ai agents this is gonna get worse, if Itch does not take action.

No, it is not normal. But it happens, and it even happens for well known engines.

And one should never disable virus scanners, especially not because the source where you downloaded something tells you so. So, yeah, it will freak out a lot of users, and rightly so. If the game is so badly coded and compiled that even the operating system scanner triggers, that is not a good sign.

Plus, there is a lot of actual malware uploaded to Itch. So people better be careful.

The problem is, that you do not have a project to link to and self promotion in forum is very limited.

You can make an announcement blog post in the posts section of your dashboard. And of course you can publish as soon as you have a prototype to show and promote this like a regular game.

Certain game engines and certain ways of creating games are prone to be detected as a virus.

Anything "clever", anything that obfuscates, runtime compressions, and any kind of internal emulations might look suspicious to a heuristic.

Godot exe runtime packing is rather bad at some versions. But just about any game engine has this in some versions. I do not know how bad the situation for python is.

Maybe a glitch. I did not notice anything strange when I had a look just now.

Older games can resurface to recent. Typically after a devlog and an update.

So I saw the typical mix of old updated games and new games in recent.

Oh, they are outdated. But old-fashioned might be a better word. There is nieches for old stuff. Even vinyl records have their audience.

And modern text-based games usually have clickable buttons and illustrations, instead text interface where you have to type your commands.

so magically ideas can stop being outdated

It is not magic. You just do not accept what outdated means and how something becomes outdated - or can become a trend again.

Text games are old-fashioned and technolgical obsolete. They are not popular. Most modern "text-based" games do feature some kind of graphics, and you will not find major game studios publishing text games. This is not an opinion or an agenda, this is an observation.

Can text-games become popular and thus in fashion again? For that, they would need to become mainstream, and that is very unlikely. Could have happened with ai chat bots, but the successful experimental games using that tech also used modern graphics.

Text-games are a nieche idea and even in that nieche, those text-games often feature illustrations and a modern user interface. Olde text games also had a text interface. The modern text games usually have a hyperlink interface.

Practices that are not done today are not outdated either, they’re simply not being done

I am not sure if you use the word outdated correctly. Things that were done in the past and are not done now, are literally outdated. That's what it means. It also means old-fashioned. Antiquated. Or simply old. Depends a bit on context. Retro would fit the video game context probably best.

Also, ideas do can get outdated. There is nothing in the definition of the meaning of "idea" that prohibits this. So I am again not sure if you use this word correctly. Is this a case of lost in translation?

An idea is only timeless in the sense that you can claim the idea exists. But you need application of the idea. Like people believing the idea or acting on it. If this goes out of fashion, the idea is outdated. Even if some people still cling to the idea or try to revive it. Outdated things can become trendy again. Especially in fashion. Though I doubt we will see ruff collars any time soon. Or big game studios releasing a pure text game.

So, your question is, how to do this or is this just in the wrong section. This is questions and support.

Genres are regular tags. The only difference is, that in the url they are written as genre- instead of tag-. But they function exactly the same as tags. A game has 1 main genre and 10 tags, You can set all of those 11 tags to be "genres". 

For randomly browsing you can use the randomizer /randomizer but it cannot filter by tags. Randomly browsing top games with the vague description of not liking visual novels is tough. A lot of games use visual novel as a tag, but are not actually visual novels. You will miss those surprising games. And a lot of games are story rich or have a lot of text and are not tagged visual novel.

So for the 1.5 million games here, it is a lot easier if you select a tag or two. Like action, or puzzle. Or whatever you actually like. Everything but visual novels is a bit vague.

Anywy, the main genre is displayed on the browse page, so they can be filtered easily client side. I made a script for that. It will not filter away all the visual novels, but at least the ones that do have it as their "main genre". I made another script (same page, no screenshot) to make games disappear, on the press of a button. This way at least you only have to look at them once.

You can also use the undocumented feature to exclude 1 tag. Pick your standard filters, add the ?... and put it in a bookmark.

/games?exclude=tg.visual-novel

If you use various filters, your selection should only be a few hundred games and you can skim most of those in a few hours.

If you use the blandest of all filters, like visual novel + horror, you have about 5k games if you look at top-rated.

/games/top-rated/genre-visual-novel/tag-comedy/tag-horror is about 300 games.

Anyway, I am not sure what question you had for community support.

Well, if you search a game about bananas, it might have the banana tag. But you will only miss it, if only ever search the banana tag.

It might still be under fruit, casual, silly, monkey, or under tags like horror, arcade, visual novel, or other tags.

The freestyle tagging on Itch is an advantage and a disadvantage. If there were a fixed pool and some guidelines which tags to use for what, there would be more tag accuracy. But people can't even properly distinguish between rogue-like and rogue-lite. There will always be disagreement over tags.

In your case, if those phone games are usually under visual novel, well, scroll past them if you spot them. I even made a tool for such things. But it only works on a desktop browser that can install tampermonkey extension.

In my opinion, those phone games should have the simulation tag. But a lot of visual novels that are not phone games also have that tag.

That's not how tags work on Itch.

The developer choses the tags. Freestyle. So there will be tags like /games/tag-phone and /games/tag-banana . And if developers think their game fits the visual novel type, they can and will chose it, even if other people would disagree. 

They might chose additional tags, but I am not aware of a popular tag that says: this game simulates a phone. Some games in the phone tag might. But tags like phone and mobile will rather be chosen to indicate that the game can run on a mobile device. 

It's the wrong category. You do not seek a collab.

And if your assets are good enough depends on the project style. They need to fit in. And it helps if the style is consistent.

People are willing to put up with all sorts of "bad" art. Look at games like Baldi's Basics. Or even at Minecraft.

Since you are broke, I am gonna assume you are not gonna pay for the commissioned assets you are asking for here.

Sooo, you are unwilling to use free assets, but you do ask for free custom made assets. 

Or is the help you are asking help on how to create the assets yourself? Your topic is ambigous. 

Can't you just look how other developer release such things? /game-assets/free/tag-16x16

I do not consider the use of AI to be unethical or a sign of a lack of imagination.

Me neither. That is why I think we need good arguments to convince any developer that might potentially use ai, to avoid it - if possible. It's just not enough to tell them they do evil or that they are lacking in skill, or are lazy.

As you said, maybe they just cannot use the tools of the trade. But they might be able to use the skills of another trade to achieve some results.

For such devs it is a choice between using ai where they have some creative input, or using stock assets, or just not have a game at all.

From that viewpoint I accept some ai in some hobby projects. But I do not tolerate it in bigger productions. Same as I would tolerate eating microwaved food at a friend's house, but not at a fancy restaurant.

What I am still trying to tell you is this: you need better arguments. The one you brought was weak and you undermined it.  You made your anti-ai argument about low effort while talking at the same time of the little effort you yourself used. You gave advice but did not use your own advice. Would you listen to someone preaching water instead of cola, if they brag about their private homebrew lemonade party?

If you use arguments that depend on opinion, you can only convince people that already share your opinion. If you cite ai as low effort to someone that spent days to create a few images that barely fit the requirement for a project, your claim does not fit their experience. And if you claim low effort to ai users that do hand edit their works, you also fail here. You are just accusing them of lazyness.

Your refusal of my points boils down to me not bolstering them with citations? Then were is your citation that ai usage equals low effort while at the same time downloading free images would not be low effort?

Where is your citation that all ai training data is stolen? Oh, wait you did cite articles. Let's see. The guardian thing is an opinion piece. The only thing this proves is that the author of the article has an opinion. So to refute my arguments you cite an opinion? Come on.

And you cite a document about a legal dispute against one specific ai model, in response to my claim that it is not generally accepted that ai art is stolen. I argued if this were true, we would see consequences of that, like a legal ban on ai.

Please read the conclusion of that document you cited. The legal claims were dismissed, except for a minor claim. And that involves basically recreating something where someone else has rights to it. And that is trivially true, as it would also not be legal do this with pen and paper.

The plea of Aleks above was better. The product will be seen as lesser 

Also your own plea just now was good: to spot Ai useage and have a negative reaction 

It explains some reasons behind: Avoid AI. People do not like it.

 You could even argue that some people think it is stolen, and therefore do not like it. But do not try to argue it is stolen. Because that ain't true in a legal sense. If potential ai users fact check this, they might just decide to use ai, since the ai opponent used a faulty argument.

Fact checking if people dislike ai yields this: https://quanticfoundry.com/2025/12/18/gen-ai/  62% very negative. Only 7% positive.

They would need a wallet system for that. And shifting money around between users is a can of worms. Taxes, refunds, and of course, bad actors.

On a platform where only users spend and only publishers earn, that is a lot easier to implement.

If Itch would handle this directly and with full wallet and user2user money flow, they probably would need an international banking license.

Only works for small games. And certain engines trying to be clever there will make your resource packed exe file trigger anti virus system.

Maybe you misunderstood. I advocate for using better arguments. 

Opinion/Argument/Advice: Avoid AI. People do not like it. You will have less customers because they do not like it.

That is clear cause and effect and one can even bolster this by citing surveys where people tell their preferences about ai.

Compare to opinion/argument: Avoid AI. It is low effort. People do not like low effort. You will have less customers because of that.

That is presented as clear cause and effect, but it is flawed. It's simply not generally true. Creating ai images yourself is not as easy as it sounds. And it can be a lot more work than just browsing for free human made images. Also, you could have hand edit your ai images. There is no category for ai images that are heavily hand edited. They are still just denoted as ai graphics.

And it gets worse since you mentioned your own method of using free assets for a contrast. If you can shrug off ai usage as low effort, so can other people shrug off your browsing for the right image as low effort. How would they know that you did not just pick the first result. And if you think selecting the right images is creative input, well, crafting a prompt and selecting the right results is a similar level of input.

In your reply just now you mentioned the stolen art argument. It's a blanket exagerration and technically not correct.  Also, you do not know if it is even true in the weak sense of unconsented training data for any particular ai system. An artist can train a network with self made art. The training data could have been 100% consented or public domain. Especially for audio you will even find things that were trained with specifically created training material. This stolen art argument is short sighted and generalizes too much. 

So my point here is: you want strong arguments against ai usage. Not weak arguments that might crumble under scrutiny or depend on personal opinion. 

That customers on average are not fans of ai is objective truth. No matter what anyone thinks about legality, lazyness or morality.

That ai usage is low effort depends highly on the developer. You do not know how lazy they were. 

That art is stolen is subjective opinion. If it were objective, ai would be outlawed and we would not have such discussions.