Editor (and overall software) quality – some priority here, please?

Talk
, ,
65

I would already be happy if simple edits would be possible without lags.
And more happy, if also navigating in the Space could be done without lags, especially without lags of 10-15 seconds after certain mouse clicks! :frowning:

And I believe that this would easy by possible.
Only problem (in my humble opinion) is, that the team hold on to a certain dream about the way how collaboration should work.
Because of that, there is all that complicated stuff running in the background, that consumes time, slows everything down and causes problems.

All that syncing and Grid-resorting and Link-renewing etc. etc. while the user does basic stuff like text edits.

Why is everything so extremely slow that the user can not even type as fast as he normally could?
We not even have WYSIWYG! – Normally essential (at least in my opinion), but we don’t have it and it seems that we will never get it (as long as no one forks the whole project and does his own thing).

I must admit that I share our “Condor Hero’s” (GuoJing) point of wiew:

Of course I hope that they get the problems into grip if they now really spent the whole summer only for bug fixing.
But to be honest, I believe the problems sit too deep inside the concept.

66

Wow, that sounds bad, fortunately I haven’t experienced any performance issues with Anytpe on my Mac so far. Maybe I’m using Anytype under different conditions than you, or maybe it’s the OS, but if I experienced what you’re describing, there would be no question for me to change the tool.

67

Yeah, and it’s even worse:
In my navigation concept I need normally 3 or 4 clicks to reach the Object I want. Let’s say 4 clicks. And a fifth to open it.
I know where everything is. Normally it should be done (to open a certain Object) in about four seconds. But in reality, it sums up to about a minute! :frowning:
– Impossible to “work” this way!

Really bad is also the writing experience in the editor.
I make many errors that I need to correct. Normally no big deal, using the arrow keys for navigating back a bit, doing the correction, and pressing arrow down for jumping back to the end.
But in reality it lags so much, that I gave up to navigate via arrow keys. Instead, I use the mouse.

There was already cases where I gave up: I wrote my text in the windows editor and did copy & paste afterwards to get it into Anytype. :frowning:

Strange is, that not everyone experiences such extreme lags.
There was some users complaining and the last 10 or 12 weeks I always thought that everyone experiences the same as me AND that the team Inevitably must know about it.
And - of course - I’ve thought that the team pauses every other work for fixing nothing else then that!
@Razor once mentioned that he can’t do more then he already did, because the problem is in the middleware.
So I’ve thought, surely the middleware team does not sleep until this problem is out of the world!
Weeks passed by, one update came after the other, but the extreme lags stayed. :frowning:

Sorry to say that, but In the current state, the product is TRASH!
No fancy new features can change this rating as long as the simplest basics are de facto not usable.
Especially I don’t give a damn about the whole collaboration feature! If that’s the thing that slows everything so much down, then better throw it away!

Same with the whole syncing. I would give it away if it helps to make the basic stuff usable.
I mean, really, why syncing at all?
The Android version was always and still is indeed trash. No widescreen for a text wring app? - Trash!!!
I don’t use that unusable app. Therefore no need to sync to a smartphone.
– Syncing with another PC, for example from desktop PC to a notebook? – Yeah, nice thing! But no need to do it in real time. I mean, to have a real time sync is fine. But I could also good live with a manual syncing after doing a mouse click at the end of the day.
– To be honest, I even prefer apps that do not the slightest internet activity on their own, until I explicit do the initial mouse click that legigimates it!

And btw.:
I never trusted any syncing to 100%.
What I would like much more is a principle where everything is inside of one big folder.
Copying that folder from one PC to another should be enough to have the other PC up to date!
(I would separate texts from media objects and do some clever other things to hold the data transfer low for such a manual copy & paste “syncing”, of course.)

OK, I wrote more then intended again …
But I’m so frustrated, what can I do?
The problem is huge but I’m not sure anymore if the team has it on the radar how huge it really is for some users?
Maybe the whole team uses high end PCs with SSDs while all the heavily suffering users use mechanical HDDs?

68

just to clarify, there are no plans to wind down anytype. we’re exploring a much clearer, more streamlined experience—but if we built it today, it would break some setups in ways far worse than what new types did. on the flip side, developing a new version would be 10x simpler, which could help solve major pain points around quality and speed of development. still, it’s too early to say if that trade-off is worth it.

the good news? anytype isn’t a cloud service, so even in a worst-case scenario, you’ll always have access to your data. thanks for your honest feedback—both the criticism and the praise.

69

that sounds rough—you definitely shouldn’t have to deal with that. it’s not normal, and it’s worth digging into the cause. are you running it from an ssd? some brave souls run it from an hdd or even a usb stick, and in those cases, slow performance is pretty much guaranteed. why suffer if you don’t have to?

70

thank you for taking the time to share this—your frustration is completely valid.

as mentioned earlier, anytype’s performance relies heavily on running from an ssd. our encryption and database setup are designed around fast storage, and there are no plans to support hdd or usb environments. every app has its requirements—this one’s non-negotiable for us.

we know that’s not ideal for everyone, and we’re genuinely sorry it’s affected your workflow so deeply. your feedback matters, and we truly appreciate how openly you’ve shared your experience.

71

WOW!
WOW, WOW, WOW!
– I quasi live in this forum, but I wasn’t aware about that!

I use it on the same PC with an HDD since 1,5 years.
OK, I was never “fast”, but it was usable without much trouble. But since 10-12 weeks, after an update, it has suddenly become so extremely slow.

There was some statements from @Razor about middleware problems and about two parallel running instances where it is a problem to shutdown one of them.
Therefore I’ve thought all the time, that the team is fully aware of the issue.

I use an HDD because they are more reliable. Also, after a worst case crash, it’s much much simpler to restore data from a defect HDD then from an SSD.
If you now say that an SSD is a requirement, then I suspect that Anytype does a lot write accesses. That’s not the ideal condition for a SSD, it shortens its lifetime.

I must say, I’m shocked!
It confirms what I’ve silently already suspected all the time: there is a huge and deep misconception in the whole thing (at least in my opinion).
– I mean, encryption and syncing etc etc. – let it away for a moment, let’s only look on the user’s writing experience:
An Atari ST from 1990 (35 years old!) with 8 MHz clock frequency and 1 MB RAM (MB, not GB!), and no HDD at all gave the user a more fluid writing experience (even with WYSIWYG!!!) then Anytype on a PC with 2 GHz clock and 16 GB RAM!

Also my 25 years old Psion Serie 5mx pro pocket computer gave me a fluid experience so that I was able to use it in a dictate in my first study. Btw. also with WYSIWYG text formatting (what Anytype hasn’t).
– I mean, really, no matter with all that encryption things and syncing things and whatever; it simply can’t be true that simple basic actions like writing in the editor need “high end” hardware to be usable at all!
There is a lot wrong with the whole conception if that’s really the case.

72

mh, I’m not “shocked”, yet I also fail to see the requirement for a high-performance SSD for a “little bit of note taking”. even with thousands of pages and an encryption layer, it feels wrong to require high-performance storage media for edits, given that you edit a page at a time, and then – i assume – write it (or a somehow encrypted version) to disk.

the only reason I can see is that to counter “known ciphertext” attacks (known plaintext?) it always re-encrypts “the whole thing”, which would be … maybe secure, yet kind of inelegant?

on the other hand, this is surely something that could be migrated over time, so … I’d say “inelegant, yet tolerable for the moment”. but i fear there are no plans to change this, at all … little cynic that I am :wink:

also that.

and @Code-Jack has a point here (atari :heart_eyes:) – encryption is not inherently slow. it’s slow compared to “no encryption” in terms of CPU cycles, yet even with the CPUs from back then the algorithm itself should be fairly fast, not even talking CPU support and GHz frequencies of today. again the only reason for “really slow” would be the re-encryption of masses of data all the time, which leads me back to my initial hunch from above.

73

You’re absolutely right that performance has changed on old machines over the past four months, and we want to clarify what’s happening under the hood—because this isn’t just a “middleware bug” or a fluke update.

We transitioned from an in-memory datastore (Badger) to an on-disk solution (SQLite) with full encryption. The reason? Reliability. SQLite is far more robust when it comes to data integrity, crashes, and long-term storage, especially with how we handle encryption and syncing. Badger, while faster in some scenarios, gave us serious trouble with corruption and data loss in edge cases—especially on less reliable systems or during abrupt shutdowns.

But here’s the tradeoff: we’re now more reliant on disk performance than before. And yes, encrypted writes and queries on spinning HDDs can be painfully slow—especially with modern apps like Anytype that juggle a lot of data behind the scenes (relations, object graphs, rich media, etc.).

So to answer your suspicion: you’re right, Anytype does a lot of writes. That’s a side effect of giving users full local-first, encrypted control over their content. SSDs are strongly recommended now—not because we love shiny things, but because HDDs simply can’t keep up with the volume and speed of encrypted reads/writes SQLite now performs.

We get that not everyone loves SSDs (especially when it comes to recovery scenarios), but modern usage patterns—do tend to lean toward SSDs out of necessity. That doesn’t mean we’re not continuing to optimize for performance where possible. But we do want to be honest that HDDs are no longer a good fit for the current architecture.

If you’re noticing a sharp slowdown after a particular release, we’re happy to dig into logs or performance metrics with you. But at the architecture level, this shift was necessary for long-term stability and trustworthiness of your data—even if it means that 1990-era WYSIWYG nostalgia now wins in a typing race.

74

Some folks misunderstood my earlier post and came away thinking we’re planning to shut down Anytype. That’s not true—and I’m sorry for the confusion my wording caused.

What I meant to say is that we’re currently struggling under the weight of legacy infrastructure and unfinished features. For context, we only just recently shut down backend systems that were still supporting the 2019 version of Anytype (!). Maintaining backwards compatibility with multiple outdated versions takes at least 10x the effort compared to building cleanly from scratch.

We’re not talking about launching a brand-new product. The idea we floated internally was to potentially create a streamlined version of the current app—same core principles, just without the baggage. In practice, that could mean removing certain legacy features and making the migration path one-way. But let me be absolutely clear: this is just one option among many, and there are no current plans to go down that road.

For now, we’re committing the next three months to fixing bugs and improving the current experience. I’ll be posting soon to gather feedback on what matters most to the community—what must be fixed, what’s most frustrating—and we’ll focus all our effort there. After that, we’ll reevaluate.

Thanks again for your patience and support. We’re not abandoning this project. Quite the opposite—we’re trying to save it from collapsing under its own weight.

75

Anytype is quite good, Anton. I, along with many other users, may not be active in the community forum, but we still use it regularly. The number of bugs is actually less than expected — at least it’s more than capable of fully replacing Apple Notes. To be honest, Notion’s editor is also full of bugs and performs surprisingly poorly, and Obsidian’s editor is a complete disaster. Sometimes, when compared with Anytype, it’s hard not to wonder how these apps built entirely with web technologies turned out so poorly. In fact, they’ve only survived thanks to their unique features.

After discussing with friends who actually use the app alongside me, we realized that there are only three core features they absolutely can’t give up: local network sync, end-to-end encryption, and web publishing. As for the rest — even if the editor isn’t perfect — it’s fine. “As long as we can type, it’s good enough.”

As someone whose life is filled with work, I can’t always fully empathize with the technical challenges you face with each task. Would it be better to build a whole new Anytype? I don’t know. And what kind of future your bold new features and wild imaginations might lead to — we don’t know that either.

But for me personally, I will always need a note-taking app with end-to-end encryption. Local network sync and database views are just pleasant surprises — and these alone mean I’ll continue to use it.

76

Thanks Anton for your clarification. I was one of them that was quite surprised about your comment. But very happy to hear you guys are commited long term.

From my side I have to add I’m fully relying on Anytype since 2022. And I have to say I absoluetely love it even though there are still some flaws here and there (for example the speed issue I keep talking about in the Nightly Ops :P).

But Anytype serves me amazingly well since many years. I use it at work, privately and even with my non-techy wife in our family-space. She loves using it as well and we are now much more organized than before.

So keep going and fill out the huge potential you have with this mighty tool! :slight_smile: :flexed_biceps:

77

Thanks @anton , this was maybe the best post you’ve ever wrote in this forum (as far as I see)!
Very well explained.

Interesting btw.:
After the step from Badger to SQLite I experienced an increase of speed in some aspects, not a decrese!
It could have to do with some other optimizations although - I don’t know.

But, as I just wrote in the other thread, Anytype does also unnecessary things that cost time.
For example, it’s not necessary after switching from one View to another, to fill the Grid that the user sees temporarily with masses of irrelevant data, only for removing them again.
And it’s also no good to refresh the Grid all the time during such things.
– Stop the refreshing until it’s ready, only then present the result to the user.
This simple “trick” alone makes it much more then 10 times faster to switch from one View to another.
.

Oh, I’m so happy that you say this! :-)=)
Why? Because I’ve often said, that every update needs information entries in the history!!!
!!!
Often we get updates with not the slightest hint in the history what has changed. :frowning:
Therefore I’m now not able to say when the bad thing has happen!
Don’t ask me when exactly I’ve installed this or that version! How could I know that after some weeks?
– This is one of the examples for the lack of communication about users often complain.

And btw. (I also say this not for the first time):
Ideal (and “normal”) would it be, if the user could see the changes in a new version BEFORE he installs it!
Anytype, unfortunately, makes it this way: “You just have pressed the trigger for the bomb you sit on”

78

I registered here just to leave this message.
Indeed, when users are satisfied with a product, they don’t leave positive feedback as often. On the contrary, when users have a negative experience, they’re much more likely to write about it.
I’ve been using Anytype for 9 months now, and I manage all my personal, household, and work-related tasks with it. It’s the best tool for my needs. I’m very satisfied with it, considering it’s still in beta and has room to grow.
I’m a UI designer and very picky about products. Many services look unfinished, and projects often ignore the details — I notice this immediately. I can confidently say that you have a great product with huge potential. Your mobile app already feels solid, and I’m sure it’ll become even more functional in the future. I understand how hard and time-consuming it is to build something, and that it takes time to refine it.

Over the past month, I’ve been testing Obsidian and was shocked at how uncritically people praise it. It feels like a developer club that built something for themselves. Some people say Anytype suffers from poor UX. I have a question for them: have you seen Obsidian, where UX/UI are practically nonexistent? I had to go through so many steps just to make things barely work and look decent.

That’s when I realized: time will pass, and Obsidian will remain a niche product for developers without a user-friendly interface, while Anytype will continue to grow and win over its audience.
As you mentioned above, now really is the time to focus on improving quality, and I’ll be waiting for that together with you.

Thanks for the product.
P.S. Sorry if it’s unclear — I translated it with AI :slight_smile:

79

I’m one of these people. Or at least I use it several times a week. I love Anytype and am using it for a lot of different note taking, planning, etc. I realize it’s buggy, but it’s in beta so I expect that.

I appreciate the UPD. It helps clarify things. As a dev I understand the need to sometimes rewrite things. I’d be interested to know what kind of baggage is holding you back. It it legacy code? Features you no longer want to support? If doing a re-write is going to make everything less buggy and future development faster and better than I’m not against it. Now would be the time to do it. I think for the most part people’s objection we be loosing their data (but it seems like you would write a one way migration plan). If you do go forward with the re-write I think it would be good as you said to have people vote on what features are most important to them (and what features they would want to see).

Is this because there are a ton of old versions you’re supporting? Trying to understand the reason you have to keep supporting those.

80

Ahah I fully quote the “have you tried” part

Obsidian feels more like a sandbox than a product. I mean, if someone prefers its approach should just switch over it.

Meanwhile Notion is needlessly chaotic and slow for everything lol

81

This captures the essence of what brought me here:

Although I’m coming from NOT utilizing any other note-taking app, I’ve been in desperate need for one. That is what started my search… I briefly signed up for Notion (specifically their new email app to start) which was probably a blessing in disguise as it seems the whole app is moving in an AI-driven context (with accompanying paywall).

I may have watched a few other overview videos about similar apps, but I was instantly sold on the idea of Anytype’s stated mission. This was further solidified when I discovered they developed the protocol behind it all; any-sync.

Am I an outlier here? The fact that Anytype was fundamentally born out of a new protocol makes considering anything else inferior. I am ready to stand corrected if there is another entity with the same philosophy, developing a modern tool that empowers MY sovereignty.

82

It’s tough I know. Every time one commits to use a SaaS one has to be prepared for a very bad situation. We have all seen what has happened to Skiff. As Anytype is also VC funded and not sustainably community backed it could vanish at anytime if the money runs out. But on the other hand Standard Notes is also community funded and it’s just rotting for 2 years by now. ugh…

I would say Anytype is also that with more to click than more to type in comparison. This is what I observed while working with both.

83

Yes, I tested Obsidian and hated the UX :grin:.
That’s also why I’m here: Anytype has great potential, but when it comes to usage (which depends on the individual and which I compare to the same usage on the other tools I use), it’s sometimes annoying. I don’t think the UX is poor, but it does have some major flaws : Labels that don’t correspond to the actual action, actions requiring too many clicks, inconsistencies in operation…
I also see everyone’s problems here, and when the same UI or UX flaw generates the same posts from newcomers over and over again… it’s tiring and confirms that this needs to be improved.
Yes, among those who don’t post on the forum, many must use it without too much trouble. And many others leave after giving it a try (several people around me were interested but after giving it a try they went elsewhere).

84

Even at the risk of repeating myself: Sharing is caring … :grimacing:

I for example would be really interested in what the team thinks about

  • … what the “core principles” are,
  • … what the “baggage” is,
  • … and which problem you are intending to solve with that hypothetical rewrite.

And just to be clear: Please do not do this; I am interesting in the thought process behind that idea, the “underlying assumptions” so to say.

Also, if you just have removed support for 2019 (!!) infrastructure, I would assume that everyone is ok with you being a lot more “move fast and break things”. Two years of backwards compatibility for a BETA app is basically eternity, if the updates automatically move things forward and there is no data loss in the process.

TL;DR: Be bolder :wink: