The team should focus solely on fixing bugs until the app reaches a stable state. New features can be added after stability is ensured. Ofc then most get dissatisfied when their feature requests won’t go though tho.
Hello @anton ,
first a suggestion:
I just used a search machine: “Anytype hardware requirements”.
But zero results.
– If Anytype really requires a SSD, I suggest to mention it on the web page where the app can be downloaded.
The main point of this post:
I was wrong when I wrote we have the problems with the low speed and unbearable lags since “10-12 weeks”.
It’s less then that, I can tell now the exact date when it suddenly has happen for the first time:
It was 2025-04-17
– Unfortunately the history doesn’t give us the information which version has which release date, therefore I can’t tell which version the update from this day has had. But I believe you are (the team is) able to find out.
I remember that I’ve presented Anytype this day to three ladies in my study group. I booted my Linux notebook computer for the presentation and opened Anytype that unexpected reacted totally unbearable slow on everything I did. That was the first time it has happen. Lags of many seconds after every single mouse click! - it was so embarrassing … 
Later the day, also my desktop PC with Windows 10 was suddenly unbelievable slow in some aspects.
For me it was clear, that the update from this day was the culprit.
But since this day, none of the following updates fixed that problem.
And I’m still not convinced about the HDD as culprit. Because:
- I’ve worked for a long time on both PCs exclusively with HDDs without problems. OK, it was never “fast”, but there wasn’t these lags in the region of multiple seconds after nearly every single mouse click.
- The biggest lag I experience if I switch in a query from a Grid that shows all Objects to another View that shows a Gallery with 36 Objects (all are Collections). It needs 15(!!!) Seconds on my windows PC and even longer on my Linux notebook!
And I can see that Anytype works not directly “smart” there: the gallery shows tons of Objects that shouldn’t appear there, before they slooowly disappear.
– This (and some other effects) can’t have to do anything with SSD vs. HDD. That’s an issue with suboptimal programming.
So, even if you insist that a SSD is a requirement, I still strongly suggest to optimize the code.
Since 2025-04-17 is Anytype on both my totally different PCs suddenly as slow as a snake, after an update. Something must have happen this day.
Anytype was always usable, although never fast. And I’m willing to believe that a SSD would help in this aspect, to make it faster, yes.
But lags in the region of some seconds after each click, even up to 15 seconds on Win 10 and even longer on the Linux machine can’t have the HDD as the only reason if it was never so extremely slow before.
@Code-Jack This is so unbelievable, 15 seconds - there must be something extremely wrong. I consider a reaction time of 3 seconds as slow and way too much!
It would be interesting to know, what complex and extensive data and queries are involved here.
I’ve never had a performance problem with Anytype, I actually find the application quite fast, faster than Coda and Notion, and I install every Anytype update that is available.
What you describe is truly amazing. What the hell happens in such incredibly long 15 seconds, one wonders. Does the app freeze? What does the CPU do? What do the log files say? Have other users had such extreme experiences?
I can tell you what happens in this extreme case, because it’s so slow that I see it.
And I will write a separate bug report for that.
Edit: here is the BR.
What happens is this:
If I switch in a Query from a Grid View with thousands of Objects (that makes no problems) to a Gallery View with my 36 Collections, it needs 15 seconds.
During these 15 seconds I see masses of unnecessary Objects appearing in the Gallery, before they become removed, quasi “one by one” - and that need time!
That means:
The Query fills the Gallery with ALL Objects AND shows them (what’s a bit stupid btw.), only then begins the filter to do his job.
Now the filter removes the unnecessary Objects “one by one” and the Gallery becomes refreshed again and again. You can watch the Objects disappear one by one.
I know this effect well from my own programs. One could call it a “beginner’s mistake”, although I don’t want to offend someone.
Filtering and resorting should happen in the background (visual refresh paused). This saves a lot time – much more time then a SSD could save compared to a HDD btw.
I love your user journey @chrsjrichards - I feel the same way about the software. After exploring many alternatives, I can definitely say that Anytype is the clear winner to build a second brain in. I’ve been trying to convert Thomas Frank’s Notion Second Brain template into Anytype, and it’s been great so far. I cannot thank the Anytype developers enough for this unique piece of software that is pretty much the lone competitor in this hugely important software space (local-first, p2p, encrypted, open code, and capable of modeling most real-life knowledge and objects).
Unfortunately, the fact that users rely on a beta software as their “Life OS” means that the developers feel pressured to maintain backwards compatibility. As a software engineer, I definitely understand the pain, and it would be hard to find a balance between dropping old versions and keeping existing users happy. I’m in favor of dropping support on older versions as long as it’s clearly communicated - it is a beta app after all. We can set an expectation that the users should keep up with the versions and announcements.
This is another hardship with open source that I fully sympathize with. As a team of engineers, the Any team needs to balance heads-down development work and putting out fires and helping out users. I definitely hope that you guys can continue focusing your time and attention pushing out great changes, but the summer bug bash sounds awesome as well. Like many other users, I also had trouble pushing it to my family and SO, but it might be a good thing since the app is in beta anyways.
This is an amazing detail to learn. Personally, I think it would be great to share these details with the community on each update, so the users can understand the potential issues. I believe the community is more technical than you might think.
Anyways, just want to give my utmost gratitude to the Anytype team. The future of this project is so bright.
Although I really love Anytype, but …
- No backlinks for images!
That makes it unusable for so many use cases that strongly depend on images! - No WYSIWYG!
You simply can’t design a form that looks in the print the same as on screen!
That makes it de facto unusable for so many important jobs in the office! - As slow as a snail!
I don’t know, how many Objects you have in your Space, but my Spaces have become unusable slow, with far less then 10.000 Objects. It’s so slow, that I can not type in normal speed, sometimes I need to wait 1-3 seconds after each single character (it’s a pain!!!) and even some seconds more after each single mouse click.
I use Anytype since 1,5 years.
One year more and I will have about 10000 Objects. What will be the speed then?
What will be in five years?
I can’t believe that changing my HDD against a SSD will do the magic that everything works normal again, without lags.
But even then - it seems we will never get WYSIWYG.
And no one knows if or when we get image-backlinks.
It’s fine that you’re so happy, but at the moment I can’t no longer share the optimism, unfortunately.
If they don’t get at least the issue with the speed solved (no matter if I use a HDD or SSD), I’ll need to look for an alternative.
– That’s so sad, because I really love Anytype! But I can’t endure the snake speed any longer, it drives me crazy every day and every time I use the program!
I hear your pain - this is probably why the software is still in beta. I’m sure the backlinks for images etc would be addressed by the team when they have cycles.
It’s also sad that this project will no longer support good performance HDD. However, when it comes to innovative local-first software, this is a tradeoff that I can understand. You have a great use case and a huge collection of objects. I think that’s amazing. We can all learn a lot from your issues that you are facing at this scale. I’d suggest you to try using an SSD. What Anton said about the database migration makes a lot of sense. A local first encrypted software will understandably need a lot of low latency IO operations, which is something that HDD is really not optimized for. If it’s at all possible, it would be amazing to test your space on an SSD device. I wouldn’t expect Anyteam to be able to offer help on every user issue since I assume that they are a small team with (probably) no significant revenue stream and a lot of pressure from VC right now.
I think the backlinks are so important that they should at least be listet in the Roadmap (even on first position!). But there is no entry.
The need for the backlinks is not new btw.
There is not even any statement that they will become implemented. - The thing with the lack of communication, we simply don’t know if or if not they become implemented.
But for WYSIWYG we have a statement from Razor. A negative one.
If they don’t change their mind, we will not get it.
So, the argument with the beta state of the software (since six years btw!!!) is not solid.
.
Yeah, I understand that.
But the migration from a RAM based database to SQLite was weeks before that update from 2025-04-17 that kicked the speed out of the program.
There are three things more that make me unhappy with that:
- I’m really extremely poor! Even if a cheep SSD costs only some Euro, it’s nearly impossible in my situation (actually it really is impossible!). And I don’t believe that it does the magic, because I believe to have proof that the software is the culprit.
- It makes a lot work to replace the hard disk even in my desktop PC, but even more in my notebook. It’s even risky to open the notebook.
And as said: I’m absolutely not convinced that it helps. It will speed up some things, no doubt. But it will never speed the 15 seconds lags down to an eye-blink, it’s the software that causes that giant lag.
I’m reluctant to do the risky and work intense (and expensive) job to open my notebook, replace the HDD in it, as well as in the desktop PC only to see that it not helps!
– At least I believe that it not helps.
A hard disk is more secure then a highly used SSD. I’m really not happy with the idea to replace a secure medium with a less secure. :-/
Much work, even risky work, cost in my situation and in the end a less secure data storage medium, only for “testing” if it helps (what I don’t believe)!?! - What also doesn’t make me happy is the fact that there was no mention about the need of a SSD (or about any other requirements) on the page where you can download Anytype.
I use the program since 1,5 years and I live in this forum, but I hear it now for the first time that a SSD is supposed to be necessary! :-/
– If that’s the case, then such a basic requirement should be mentioned on the download-page.
Yes, even more, if a user nevertheless tries to install the program, the installer should give an alarm that the requirement is not fulfilled and the installation makes no sense.
All this wasn’t the case.
I’ve used Anytype all the time since 1,5 years. Now I have my some thousand Objects in it (and it’s “not easy” to get them exported out, btw.!) and now I’m confronted with the statement that my PCs aren’t (and wasn’t never) good enough?
At the moment I really doubt a lot …
(I have some more hard critic points, but they don’t belong into this thread.)
for backlinks you mean backlinks to images?
I mean, if I have a query for images and I open one of the images shown there, I urgently need to find the Pages that use the image in it!
Let’s say, I have a database about flowers.
For each flower I make a Page that describes the flower and has an image in it.
Later (days or weeks or months later) I scroll through my query for images and I see a very interesting flower there.
I want to know more about it, but where is the page about it???
– I gave this example for flowers, but in reality I need it in my study for illnesses.
I have masses of images about skin illnesses, but it makes no sense to put them into Anytype, because I’m never able to find the Page that belongs to the image.
Edit: an example:
Without Backlinks to the Pages: wtf does the images show?
I already gave each of them a unique name, but this causes an immense lot of work and it forces me to use the global search to (hopefully) find the Page that uses a specific image.
– That’s too much friction and much to cumbersome to live work this way!
Imagine you sit in the study while the professor teaches and you want to quickly look for something …
Unfortunately, I don’t quite understand the desire for backlinks for objects like images. Assuming an image is used on 1000 different pages - what good would 1000 backlinks on 1000 pages that use the image do me?
I would rather like Anytype to finally make it possible (as it has been implemented in Notion for ages) to pull the cover image of gallery cards from the content of a page.
Since in Anytype you can only upload images via pages (or other types), an image is always on a page (or another type) first anyway, so why should I query images separately as objects and then get backlinks to the pages that use the corresponding image?
I get around the missing feature of showing page content as cover on cards by using a custom property of the type “Image” for each page, which I can then display as a cover image for cards in gallery view. I name that property “thumbnail image”. The additional manual work involved in searching for and selecting the image for the thumbnail image property is of course enormous with larger amounts of data.
Notion can do this automatically - the first image on a page is displayed as a thumbnail for the cards in gallery view, regardless of the rest of the content.
But perhaps I have misunderstood your wish for backlinks for images, or the way you would ideally like to work with images?
First, this is (for me) an unrealistic case.
But even then, we already have some Object and/or Collections, where is is in deed realistic that they have not 1000, but maybe 300 backlinks. And it works, what’s the problem with so many backlinks?
For images I have mostly only one, or maybe two or three backlinks.
I also miss this feature.
It would be a “nice to have”, yes, But it’s not nearly as important as the missing backlink feature.
An example:
A patient has a skin illness. I don’t have a clue what he has, therefore I look into the Query for images to find a similar looking image.
I’ll find there an image that fits. But how can I find the Page that describes the illness?
Normally all images have a name like “image” or something else that doesn’t tell you anything. It’s impossible to find the page then!
For this reason I give each image a unique name, so that I can use the global search.
But this is way too cumbersome.
Try it: Upload 10 images and give each of them a unique name!
You’ll see how cumbersome it is, to open each of them as object and rename them! Alone this simple task costs much too much time to live with it.
But then, in addition, the need to use global search each time to find the Page that uses the images is in practice simply not realistic.
If you don’t believe it, simply try it! – It causes WAY too much friction!
A simple mouse click on the backlink could and should do the job!
I’m with you. The missing card feature would be very helpful, yes.
But in practice I have in most cases more then only one image per Page, therefore the card feature alone is not a sufficient replacement for the image-Query with backlinks.
I have (and everyone of us has) a lot images where I’m not even sure if they are already obsolete and can become deleted.
And I can’t find out, because of the missing backlinks.
Such images are often left overs from Pages that are already deleted, but the images stay and are still appearing in the Query for images.
I can’t clean up the garbage.
Interesting take. In the past 20 years, you didn’t notice an increase in interest for privacy on the internet? I saw a huge increase in interest for privacy in just the past few years alone. Sure people still have Facebook/Insta and other socials, but they are certainly way more cognizant of what they share to the public nowadays than they were just 10 years ago. I use to remember Snapchat having a feature where everyone that follows you can see your live location and people would just gladly show where they are back then, not anymore though.
Proton is growing with more and more users everyday, Apple adopted Advance Data Protection to make data E2EE, Signal, Brave, DuckDuckGo, Mullvad, more and more softwares are appearing being pushed with privacy at the forefront, and so many examples I can draw from.
Yes, the privacy number is still small compared to the general population, but it’s growing rapidly. But the privacy market is big enough for all the softwares out there. I actually don’t think there’s enough privacy centric softwares. The people that start being privacy conscious won’t really do back to stop being that way, it’s not like a habit like going to the gym imo. Once they get on the DeGoogle train for example, they very likely won’t be moving from Filen back to Google Drive. So the number will only be growing; not dropping, not stagnant, but growing.
And then we have situations like with Snowden, recently with the UK trying to get Apple to build a backdoor to their E2EE to invade people’s privacy, and many more in between; we’ll definitely see a lot more of these in the future. And my take is that the more of these happen, the more it’s going to drive people to favor privacy, pushing new people into the privacy wagon and adding to that growing number exponentially.
I think there’s a mix up with ownership vs privacy with the Switch example. FYI I have a Steam account, but I won’t be storing my photos in Google Photos.
There will always be people that will still post their whole life online and those platforms you mention will still be around, no doubt about that; but this doesn’t take away from the number of people with interest in privacy alternatives growing day by day. The privacy market is getting bigger, and that’s the only direction it’s heading.
However, I appreciate the input that I read, I love seeing other’s perspective on privacy.
Thank you for clarifying, Anton.
please let’s not do any „not my use case, so you might be wrong“ remarks. i hit that limitation on day 2, and he gave multiple explanations. if in doubt the team could cast an „official vote“ and communicate the result. also, EVERYTHING has backlinks, it just makes SENSE.
i think that was never the intended use case for anytype: a second brain is information organising, not page layout. and the latter we do have in rudimentary form, and your use case could maybe be mitigated (maybe only to a degree) by plugins.
The is an “official vote” here:
most people just use whatsapp and just roll their eyes when people who are aware point them to signal or something similar. 20% might even install and try signal, see no one is really on there from their contact list and just leave. I also noticed as I was the only person why people in my surroundings installed signal, they didn’t open it for weeks or a few months (we didn’t have regular contact, just from time to time) and Signal stopped sending notifications. Then they all forgot about they even installed Signal. Just no one used it (other than me) with them.
Or Instagram, the population on Insta for example gets younger and younger. A lot of parents might be aware and more conscious than previously but a lot of them are not. Influencers target even the youngest nowadays and there is no awareness of privacy whatsoever.
Sure privacy companies grow a lot but this is not because of the general public. I my rural surroundings I would know no one who even knows about Apple’s E2EE or Proton. As long as Apple doesn’t make it the default (it’s opt in) all those people won’t use it because it is just inconvenient for them.
@_sky unfortunately I think @krst is right.
I would say the high time for privacy was in the three years before Snowden.
In Germany we have had the “Piratenpartei”, a political movement for more privacy and data security.
They already got more then 10 percent in some elections, I believe 14% was their best result.
I remember a very successful petition against a proposed, very bad internet law from “Zensursula” (Ursula v.d. Leyen).
(In my opinion the most evil person in the whole politic, btw.!)
Then came Snowden. I celebrate that and I’ve thought, the movement will now grow through the ceiling!
But unfortunately it seems, Snowden was the point where most people who cared about such topics gave up.
The Piratenpartei suddenly became insignificant and lost lost all elections since then.
Soon they became totally irrelevant in the politic, till today.
There were a few political demonstrations for more data protection, but they have had ridiculously few participants.
I couldn’t believe it, but the CDU, who was all the time the most evil enemy for the privacy movement, won election after election, with only one exception (where we got the “Ample”, but they was not much better in this aspect). The Piratenpartei is totally out, no one speaks about them anymore.
That means, the privacy movement has lost fatally!
Most people simply don’t care.
Since Snowden, they have seemingly given up in resignation.
Of course I think that privacy is very important!
Some others still think so.
But you can’t say that the movement is growing - the opposite has happen, that’s totally clear.
Even Anytype makes in my opinion some mistakes that I merely can believe.
For example the highly discussed fact that hovering the mouse over a web link gathers some information from the target. I still see it as a big problem in specific cases, but the team thinks otherwise.
As a workaround I paste many web links as text instead as links. If that wouldn’t be possible I would have gone one year ago!
A second example:
Try to buy a membership plan.
You are forced do give payment providers like Stripe a lot of your data.
Your name & address, your email address and even your real Anytype-Id!
– I call this combination “toxic”!
The Anytype-Id stands here open in each bug report. That means, that someone who has all these data is able to make the connection to my account here and to everything I ever wrote here in the forum!
– I don’t like that!!!
- Why on earth does our Anytype-Id appear visible for everyone in each bug report?
Bad enough to ask for such data, but why even showing them open for everyone and for every search robot? - And why on earth needs also Stripe get the real Anytype-Id, instead of an anonymous code?
It would be possible to pay a membership plan in the same way as paying credit for a prepaid SIM-card.
But for unknown reasons, the team didn’t implement such a privacy friendly was to pay.
I received from Anytype a voucher code because of my contribution. But in fact I’m still reluctant to take that friendly offer, because I simply don’t want to give Stripe my data. And I have my good reasons why I don’t want it.
These data would appear also on my bank’s account statement. Therefore also the tax office and (even worse) also a specific other authority (that has unlimited access to my bank’s account statement) would see it there and could use the Anytype-Id to easily find me here in the forum, that I for good reasons not want.
I really wonder why Anytype does such things if they really care about privacy?
@anton once scolded me because he thought that I somehow mix up privacy and data security.
Seemingly he didn’t understand my point while I didn’t understood his point. We couldn’t find a consensus.
In my opinion, security and privacy are connected to each other. It’s a thematic complex.
Alone that we have had a big discussion about things like that shows me, that many people (including the team) doesn’t care enough about privacy.
For me it was all the time (and still is) already a bit too much to feel comfortable to post my Anytype-Id in each bug report. But at least I’m (more or less) “anonymous” here. Although I dox myself more and more with every screenshot and every video.
But on top of that, giving Stripe the missing parts of the puzzle is really too much for me.
In my opinion, no one who has paid for a membership plan, cares very much about privacy.
– We are a bit off topic now, but I find it important to say all this.
Do you use Anytype on your phone? I’m sure your phone has an SSD.
I think even if privacy isn’t the top concern for some users, the security and performance of a local first and encrypted software are enough to convince most people to migrate away from notion.
Nope, I don’t use it on a phone.
I’m not very tech-savvy, just curious — if I use the stablecoin USDT to pay for a membership, will my information still be exposed?