Community Q&A - December Town Hall 2025

Anton: Of course, there were many interesting moments, but we remember better the latest moment when I received the first one-to-one message in Anytype. I felt really well, because this is software I love, I use every day, and I received this direct communication channel that nobody in the world can break and take from me. So it was a very wonderful feeling.

Another wonderful moment was when Kaye joined our team. When Kaye got that offer and joined our team, I think it was another moment. We had a lot, but I think these two are the latest and the happiest for me.

Anton: In general, the roadmap we outlined today, we are super excited. And of course, as a business, we are super excited to reach social sustainability as soon as possible, so we can be completely independent. We believe if we deliver on the roadmap, maybe we'll be there. Since six years, not many things changed in terms of excitement. We want to build this software and that many people would use it.

Kaye: When I first joined the team at Anytype, what was clear was that this year there had really just been this deep reflective moment about truly understanding what this greater vision was, what the company and product strategy were, and what that meant for the long-term future. What was it going to take to be successful in really building this kind of sustainable network of local-first architecture, applications, and users, all interacting permissionlessly, decentralized, and independently of any kind of centralized control.

Kaye: Let me answer this super directly and be upfront: Anytype is not profitable. Like most growing startups, we're still in investment mode, focusing on building the best possible infrastructure and product while growing the community at the same time. Obviously, this is going to be a hard journey. If anyone here has worked in a startup or founded their own startup, you know how tough it is, and we don't want to be sugar-coating that in any way.

However, there's a very clear path forward for us. What we're really even more excited about is the fact that there's a really core group of very supportive investors and community members who really believe in what we're doing, and that makes things much easier to forge ahead.

Anton: From our perspective, if each power user would buy a subscription right now, then Anytype would be almost profitable. So there's a path forward.

Another point: whatever would happen (we don't know the future), but what's beauty in our software is that the code is open, your software is local, so whatever you're investing, there is no chance you'll lose it.

Kaye: Believe me, I experienced the problems with search every day, so we are aware. We groomed this project today a little bit. We have it on the roadmap to fix. There is a long list we're compiling to start working on it.

Anton: Unfortunately, there is no answer yet for this. We have 1.5 people on Android right now. Android and iOS are keeping up to desktop. If we had more paying members, then we would invest more, but right now, no promise. It may come together with List 2.0.

The same answer applies - we can't commit to a certain time, but obviously Kanban on all platforms will have feature parity eventually.

Anton: It's possible to adjust, but it's definitely not on top of our roadmap. You can find on the forum a post about custom CSS, but I understand that's not exactly what you're asking about. No plans for that right now. When Block Editor 2.0 gets done, we can revisit a lot of these conversations.

Anton: I think we can do it. Should be not hard to do. If you could point out on the feature request which already exists, we can look into it. Unfortunately, there are so many feature requests that we can't go through all of them precisely - we just don't have capacity for that. You can speed things up by directly sharing it in the community chat.

Anton: We have it, but not on the user level. You can manually add CSS - many people do it. We have a post on our forum that's quite popular where you can find instructions. But we don't plan to add native user customizability on CSS today.

Kaye: This is a huge project for us, and absolutely something that we are actively addressing and working on. It's a very high-level and strategic question, and it's really much more complicated than it seems because of everything being local-first and end-to-end encrypted. This is definitely something we're working on, and we do want to provide a better system for people to be able to do permissions in the future, 100%.

Anton: What's coming soon is admin role and change of ownership. Then also some limitations on who can do what within the space. Regarding the whole system, there is a challenge not only connected to our cryptography, but in general conceptually, because it's a graph of data within your channel or space. In other software like Notion, you have hierarchy, and it's really easy to work with hierarchy in access control lists. But in our system, we have types, everything is interconnected. Even from the user perspective standpoint, making it work is a big project and important for us.

Anton: I don't think it's influenced somehow. It was easy to implement, so Razor did it. Handwriting is another scope of project. We're not sure it's a priority for now.

Handwriting in general, or some sort of sketching tool in Anytype in general to be able to do that, is not something that's actively being discussed today.

Kaye: For those who are not aware, there are a lot of chat control regulations happening across the EU. Essentially, what they're trying to do is to install some sort of backdoor for all chat and messaging services in order for companies to be able to provide their users' access data to authorities. The agenda is child pornography—obviously, when that's happening, they want to be able to go into people's devices and catch all the bad actors.

I can't speak for everyone. This is a very personal and political issue. I can say that I think this is a very bad move by governments to be pushing for this type of direction, because it deeply infringes on privacy rights for a lot of folks, and there's really big ramifications.

The analogy I typically use is a kitchen knife. At home, you have a kitchen knife, and you use it for many productive things. But somebody could use that for domestic violence and murder somebody. Does that mean the government should ban or remove all abilities to have kitchen knives? There's a very dangerous thing that happens when there's this level of intervention within very important primitives of the way societies function.

It's definitely worrying that these regulations are being discussed. We, as a startup, have no real influence of this because we're just too small. But it is definitely something we are watching from the sidelines.

Anton: I just want to say that our technology does not allow this kind of behavior. So probably we would not officially work in these countries where it's not allowed.

Kaye: Just to be clear, even if we wanted to comply, we wouldn't be able to because the technology does not enable this. It is all end-to-end encrypted, so there is full privacy for all of our users.

Kaye: I heard Switzerland has a similar law currently being floated, but they have to do a very democratic process to get that through. So it's all very complicated, but there are laws all talking about chat privacy across much of the EU at the moment.

Anton: It's in general a huge topic for us - authorization. We were looking into different solutions like Web3. We definitely need to implement it. Probably the end of next year. But definitely it would be a simple login like you use Google, then you have your password like Shamir's Secret Sharing scheme. But it's a bit challenging in our case.

Kaye: Just for folks to understand, making things permissionless makes it harder than just doing things the simple cloud platform way.

Kaye: There will definitely be a return of more regular town halls. As Anton mentioned at the beginning of this call, Zhanna has been on maternity leave, and Anton has had to shoulder a lot of work in her absence. So it's just been a little bit tougher for pure human labor reasons to do that. But 100% moving forward, there should be more regular ones. I can definitely hard commit that there'll be at least three next year. We're thinking about a cadence of maybe bi-monthly or quarterly, and we'll see how things land, but definitely there'll be more town halls.

Kaye: Anton really answered this with List 2.0. Essentially, when we move lists, collections, queries, sets - all those things as you guys know it - onto the new AnyStore infrastructure, and we re-architect how we believe that thing should work with tags and all the rest, the entire listing system will be better than it is today. So you can expect very big improvements next year.

Anton: Not just improvements - really like a paradigm shift.

Anton: There are some limitations in API, but it will be developed further. We have developers.anytype.io where you can see documentation. We even have official MCP for Anytype and Raycast extension that Jannis kindly developed.

Regarding version 4, it's currently ready for review, and we're finishing the last things. It should be deployed until the end of December.

Roma: Yes, I think it's already providing the local URLs for the files - you can download them locally. I think it was already in Stage 3 API, but maybe I'm wrong here. I will write this down, and in the Stage 4 API, we will put this in the documentation.

Anton: It's still in beta right now, and we didn't make any research regarding changes we need to make. It's a community element, so we'll be very happy to accept your contributions.

Anton: This will be part of the List 2.0 project.

Anton: We were discussing this internally. Unfortunately, right now it's not a priority. But I think it's a clear use case. It depends if we bet on this kind of use case. Right now maybe not a priority, but if it's a project that's quite short, we can think about this. I agree, sometimes when you publish a picture or something and you ask people in the space, you wonder if somebody even looked at it or not. It makes sense.

Anton: Everything from scratch doesn't make sense in our case. A lot of our business logic is reused on clients because it's written in Go. In terms of Flutter, I've never seen really great Flutter applications on mobile devices. Would be great if you can share these great examples - I tried to find but never found.

Kaye: Multi-object publishing/web publishing is something we want to work on in the future, but there is no ETA at the moment.

Anton: We started work on it as a heavy project, but right now we're focusing on another service a little bit. But definitely we'll deliver it. We just can't answer the precise question right now.

Kaye: Absolutely, we could not agree more. It's definitely something we're considering, but we just have no timeline on it at the present. Fully being able to publish multiple objects including queries and collections is something we understand the value of.

Anton: It's technically possible even right now. We just don't have a UI for this in the application, and it didn't have enough user requests for this.

Anton: To be honest, we have really scarce resources, and spending them on building for self-hosters publishing means we would not deliver some features for people who are paying. The code is open - you can build it yourself. In the future when we have more resources, we can do it. We're not doing it not because we want to lock it, but just because it's not the right thing to put our efforts in.

Anton: Yeah, we are working on them. We are ready - we have designs, we groomed them. We're prioritizing at the end of this year.

Roma: Very good question. This is one reason we are not hurrying to deliver AI features. Ultimately, we are targeting local models and private AI computing. In the meantime, we want to provide BYOK (Bring Your Own Key) for users so they can take responsibility if they feel comfortable using some cloud or self-hosted models.

Kaye: It being local, it being self-hosted, there's many different ways that we can do it. Confidential Compute is another way. There's a lot of different ways we're thinking about that. There's not really only just one solution, and actually providing users with a choice is a key part of it.

Anton: We don't plan to make our own models. That's your choice - you can bring whatever model you want. We unfortunately have no ability to block you from using a model of your choice. All these limitations reside in the model itself.

But if the question is about whether there would be limitations on which objects the model can access in Anytype - that's for sure. There will be explicit UI where you can choose what you can share, what you cannot.

Anton: When we're prioritizing roadmap, we're looking into how hard it is to implement. If you're building on cloud, building something like reminders is just one day of work, maybe two days of work, and then two weeks of testing. In our case, reminders in distributed systems are quite a pain. For example, notifications - we spent maybe half a year building these notifications so they're still private. This is why when we are prioritizing, we are looking into how hard it is to implement.

Anton: We prioritize based on the impact and how hard it would be to develop. Quality of life and bugs always come to each release, but the big projects we prioritize based on this and also strategy.

Anton: No plans for this yet. It's out of scope right now because obviously there are many other things that need to be fixed before that. I wouldn't say it's in the next two years even.

Kaye: I think Sergey has mentioned that we might be putting some add-ons where you can add on certain parts to your plan to increase things, but we'll look into this. It's certainly something that some folks on the team are thinking about.

Anton: Yeah, when we release self-hosting, we can consider this. It would be an option, definitely.

Anton: Actually, we think the current plan is kind of a family plan, because if you buy an account, then whatever spaces you create, I think up to 3 people can use them. All limits only apply to the owner, not to the other people. So basically, if you create several spaces, then your family can use it.

Anton: No, not yet. We need to discuss precisely.

Roma: Our current direction we want to go is that you can run some really easy, light node of Anytype on your machine, and it will be part of the network. It will store all your files, everything. Basically, we're going to deliver this auto-download feature. If you run the desktop client on your computer 24/7 online, then it will be kind of your backup, and it will also be able to communicate in a peer-to-peer network of your local network.

I think this is one of the ideas, because the current self-hosting is kind of complicated for people to set up at their home. We want this experience to be much easier.

Kaye: At the moment, Evernote is relatively low on our priority. Not many people are coming from Evernote, so there's no active plans at the moment, unfortunately.

Anton: But I know there are some exporters from Evernote to Markdown, so you can export to Markdown and then import to Anytype.

Kaye: We wish that we could easily do that. As a bigger topic, yes, importing will always be something we want to do well and focus on, but it isn't easy. But absolutely it's always something that we want to improve - coming from other platforms to Anytype.

Anton: Yeah, we are right now exactly working on that. We're facing this problem ourselves every day, and we're going to fix it very, very soon.

Kaye: As everybody who's aware of using Anytype, once you're in a space, you can't really do sharing of specific objects. That goes into our entire topic we've talked about, which is permissions. How do we do permissions, and how do we do that in a way that makes sense? That's something we're actively working on these days.

Kaye: Right now, if you copy an object link and you post it in another space, and then you click it and you have access to the other space, then you just jump straight to that space. So that technically works now.

Anton: Definitely, it's very essential. For this, we would need to rework many parts of the application to make it keyboard-first.

Kaye: I do think we want to improve the way that we solicit feedback from the community. It's definitely something we're working on next year. Not to go too much behind the hood, but we actually need to improve how we as a team handle data and insights and feedback, and all these types of things, aggregate them, and then have those things affect our prioritization and our roadmap. So it's something we'll definitely work on in the greater course of just trying to be better with receiving feedback in a more streamlined way.

Anton: Yeah, it's definitely cool to have it. Right now, we don't plan to do it. Let's fix basics first. But cool thing that with API, and then we plan to release Headless Client also, which basically can run Anytype always on some machine, and you can connect to end-to-end service and use other services.

Kaye: No lock-in is a really important philosophy because it's a way to make sure that you are actually having autonomy. No lock-in can mean multiple things. It can mean open formats, like the fact that Anytype uses Markdown. It can be the fact that you have the ability to self-host versus using a network provider, or syncing via peer-to-peer. There's a lot of different aspects.

But the essential goal here is that whenever you're using Anytype, you should not feel like all of your data, all of yourself, is locked into the system and you can't get it out. Or you feel like the vendor is really preventing you from leaving. That's really the spirit of what Anytype strives to do, and you can see it on a technical level.

Kaye: Just to give folks the story, I'm a hardcore user of Notion and Slack, and I've pretty much completely been using Anytype. I use it for really everything. We can see so many different ways of using Anytype for a lot of different workflows. Actually, most recently, because we do a lot of stuff with our investors, we built our entire data room inside Anytype as well, which was super cool and really fun. We can chat with people directly about all of our strategies and our roadmaps and share things very uniquely in a really cool and safe, private way. So we are definitely using Anytype in unique ways, and we're always excited to see how other people are using it as well.

Anton: For task management as a team, we don't use it because it's not ready for that. So for development, we use Linear. Someday we'll move to Anytype. Right now, we moved from Slack for knowledge plus discussions. When it's ready, we have a list of things that we need to finish before we can move from Linear, because Linear obviously is a great software tool.

Kaye: It's really not very useful for us just to throw random numbers out there without any context. What I'd rather do is tell you guys a very transparent and clear story about our metrics.

Our metrics right now tell us that the power users who get Anytype, who really understand Anytype, and who really use Anytype, we have really pretty incredible metrics. They really use Anytype not only for a long period of time, but they even increase their usage over time. When you look at the health of a product or a business, one of the most important trends you want to look at is what we call a cohort retention curve. When you see that amongst power users, it actually goes up over time - it's a really, really healthy and positive signal.

Now, that's one good thing. One of the really bad things that we have, to be transparent, is our activation. Most people come to Anytype, and they're just like, "Okay, I don't get it". That's something we really need to work on - making it way easier to understand Anytype, to use it, and get to the point where you see the benefits and value of it. That's why we're working heavily next year on stuff like Anytype Learn, the new onboarding, and more tutorials and all these types of things, because that's really a big hindrance for us at this point.

Thank you to everyone who join!