Anytwo and the future of Anytype

News
1

Thank you to everyone who joined us for our latest Town Hall and apologies to everybody for the technical difficulties. We’re truly sorry about that and will do better next time. Below is a recap of the major announcements and what’s next. But it’s best to watch the Town Hall in its entirety, especially from the What We’ve Been Cooking section.

What We’ve Been Cooking

We have completely re-worked the foundations of the application to address long-standing issues and make the product more stable, fast, and future-proof.

New Foundations

  • New Infrastructure: The legacy, complex Middleware layer has been completely replaced with the our new Any Database. This vastly improves app stability, reduces regression bugs, and makes for much faster app development. Additionally, the Files Node was rebuilt from scratch for faster uploads/downloads, reliable file sync, accurate space recovery on deletion, and lower self-hosting costs. We’ve also added global peer-to-peer sync in our infrastructure, eliminating the need to share a local network.
  • New Editor: Abandoned the old block-DAG editor that was the oldest part of the Anytype product which was built 6 years ago. We’ve now adopted the future-proof open source text editor Lexical, which is used on Instagram, Whatsapp, etc. The new editor uses real Markdown, bringing instant responsiveness, and solves many long-standing issues.
  • New Organisation System: Implements a revamped Types system, generally known as Collections 2.0 — along with a completely new way to organise your objects.
  • New Space System: Introduced a intermediate layer between your vault and individual spaces that acts as a unified hub. This allows you to share properties, templates, and assets across spaces and create aggregated cross-space views.
  • New Search: Completely rebuilt search infrastructure to support reliable full-text and vector search across all your spaces at once.
  • New Integrations: Added a brand-new integration system built to easily move data in and out of external apps, such as calendars, email, and meeting recorders.

In summary: It’s more reliable. It’s more performant. It’s more extensible. It’s more integrated with other apps. It’s better in almost every way. The result is an application that feels like it’s your very own custom software that’s connected to all parts of your life

Migration from Anytype

In pursuit of a smoother product experience and faster development speed, we have sacrificed backwards compatibility from Anytype. In other words, if you want to use the new build, you will need to commit to a hard migration from Anytype. We will build a bridge from Anytype to the new build to make the migration as smooth as we can.

Anytwo

Anytwo has taken all of the learnings from the last 7 years of Anytype and has been reimagined for the future of knowledge and software. Many long-standing feature requests and bugs are no longer blocked by legacy code. It’s functional today, and we will invite the first round of testers next week. Please watch the demo video for more information.

  • Editor: Synced blocks, block links, word count, table text styling, table formulas, image captions, instant markdown support, and more.
  • Organisation: Collections 2.0, richer back links, side-peek and groups on Grid layout, and more.
  • Mini-apps: Deep customisation of your spaces with new capabilities, such as a Wiki, Journal, Contact Book, and more.
  • Integrations: Easily get your data in and out of your most used software—generate any connector on demand.
  • Agentic Capabilities: Work with your knowledge base and integrations alongside an agent powered by a model of your choice. Turn off AI button also exists for those who do not plan to use AI.

The outline of what will happen next is likely to change, however here are the high level points:

  1. We are opening to initial testers next week. Sign up here: Any: Early Tester Waitlist
  2. Testers will be able to greatly influence feature development, roadmap, and prioritisation.
  3. We plan to release the product in its first iteration publicly by the end of the year. It will not have feature parity to Anytype in its first releases, however it may have enough depending on your individual needs.
  4. We have not decided on the membership model for Anytwo yet, however it will have a free tier but will be structured differently to Anytype.

AnytwoDB

Our new infrastructure replaces our legacy, debt-heavy Middleware with a clean, product-agnostic database layer. In short, we now have the world’s best-in-class tool to build local-first software with high quality and faster development.

This empowers vibe-coders and developers who want to build software that’s local-first out of the box and simply focus on product development rather than infrastructure. Learn more at docs.any.org and if you’re interested in using it, please reach out to Kaye directly. We will set you up on a call with our team.

The significance of AnytwoDB is that provides us, and other app creators, an open-ended canvas to build products that can easily integrate together. Anytwo is built on AnytwoDB, as well as Anyscribe.

Anyscribe

As a weekend and moonlight project, a non-coder and designer on our team built a local meeting recorder application—we call it Anyscribe. You can record meetings and get transcriptions locally on device using your choice of model. We will release this in the future for free to our users, who can integrate it directly with Anytwo to get your meeting transcriptions and notes directly into your knowledge base. We hope this application is a showcase of the power of AnytwoDB.

What happens to Anytype

Anytype will move from Source Available to fully Open Source under the MIT License. We will continue to support Anytype as much as the community does. And your continued support directly helps us to keep the lights on and build tools that you love to use.

  • You can continue to use Anytype forever, it’s not going anywhere.
  • Anytype’s infrastructure is fully sustainable today thanks for 2% of the paying users of the community. If nothing changes, it remains up indefinitely for both free and paid users.
  • Development pace will slow down because we’re focusing on Anytwo, however more memberships to Anytype can increase releases and bug fixes. Rollups and better search is coming up next.
  • If Anytype matters to you, please consider supporting the project through a membership. :heart:

Memberships

We have not made any decisions on how we will treat memberships/perks from users on Anytype moving to Anytwo, however we believe we will be more than fair to all of our community members who have supported us over the years. Once we have a clearer idea, we will certainly share specifics. Regardless, the advent of Anytwo does not impact your current usage of Anytype whatsoever—you can continue as you have been.

We plan to release lifetime memberships to Anytype which will transfer over to Anytwo once it releases, however we have yet to finalise it.

Anytype Business

Our core mission has always been anchored in digital freedom, data sovereignty, and privacy by default. However, as organisations have attempted to adopt Anytype, we found ourselves at odds with how we want to protect individuals on Anytype.

  • Sovereignty vs. Liability: Personal privacy on Anytype means encryption where only the user holds the key. For a business, if an employee loses their vault key containing regulated customer data, this can be legal liability issues in data management. On cloud applications, administrative recovery is easy because they hold all the keys. On Anytype, it isn’t because only the user holds the key.
  • The Solution: Rather than compromising our core values or forcing corporate requirements onto individual users, we are introducing a dedicated codebase and license tailored specifically for organizations—Anytype for Business.
  • Shared Foundations: The business codebase shares 99% of its foundation with the personal version. Core developments will flow bidirectionally between both products. This isn’t about gate-keeping features, it’s about separating what doesn’t make sense for individuals to the business version of Anytype.
  • Takeaway: For teams and organisations interested in using Anytype in their workplace, please visit business.anytype.io for more details. For all personal users, this doesn’t impact your experience.

Community Feedback

In combination with the Community Roadmap Survey and AI Ally Tester Experience, we have learned a lot about what our community needs. There were some hard learnings and reality checks. At a high level:

  • 80% of our community is still waiting for new features. And they want them faster. Fixing bugs, making incremental product improvements, and building a stable product is desired, but it’s just not good enough. We need to make bigger changes.
  • Although there is controversy surrounding AI, overall there is real demand for this functionality to exist in the community—it’s matter of maintaining our strong standards on sovereignty, security, and privacy.

AI in Software

Much of our community is not fully aware of the nuances when it comes to what AI in software means. We elaborate our stance and vision on it as follows:

  • Local-models on powerful devices paired with local-first software is the future. This gives everybody the power of AI with virtually no sacrifice to their privacy, security, and sovereignty. And today, this is already a reality.
  • We believe the flow should be: First, choose if you want to use AI or not. Second, choose which spaces you want to use it in. Third, choose what models.

Coming up on Anytype

Expect Rollups, improved Search, and Formulas as the primary area of focus. Additionally, we’re making massive improvements to:

  • API v2 — allows for much manipulation of objects, views, and tools in Anytype. Whatever you can do as a user should be accessible through the API.
  • Any-block v2 — allows for better data portability and compatibility for your content in Anytype. This reduces your feeling of lock-in that what is in Anytype can only stay there. You should always feel that you’re in control of your data.

The Anytype roadmap will be properly updated after the dust settles with the first testing phase of Anytwo. However you can be sure to expect the above items as being part of the next releases.

5

Note to our community: I want to say that the Anytype community has been incredibly supportive over the years and has picked us up on the many days we have been down. We know that this announcement brings both excitement and disappointment because of the need to migrate from Anytype (no backwards compatibility). Believe me when I say that this was not an easy decision and has been deliberated on for a very long time (years, in fact). We’ve made this decision based on so many factors, but we ultimately hope that it’s the right one for getting you an amazing tool to empower your digital freedoms.

Please don’t be shy in sharing your thoughts, both positive and negative. We’re always here to listen and nothing is ever set in stone. We depend on you, our community, and are not unaware of that fact. Thanks again to all of the contributors to Anytype, you are incredible humans with generous hearts. From bug testers, to content creators. From translators, to forum debaters. Your efforts are not wasted on us, even if it can feel that way at times.

We are grateful to have this global community.

6

I wanted to once again thank the team for their effort and an amazing demo/presentation. I’m not sure what some of the people here have seen, but I think from anywhere you look at it, it was absolutely mind-blowing. :exploding_head:

The way I see it, we’re going to get a revolutionary product in Anytwo which because of its underlying structure, will enable so many of our feature requests along with, according to the team, an elevated pace of development in the future. Trust me, you guys are not ready for what you’re about to see. Just go and watch the recording from the town hall. It was ‘one more thing’ on top of another ‘one more thing’. :joy:

As for the pricing, quite frankly, I’m amazed as to how until this date we have access to this many features for free in Anytype. When we look at direct competitors such as Notion and Capacities, they have locked much more of their features behind the paywall compared to Anytype. But, in any case, it will come down to whether Anytwo would offer enough compelling and unique features for any individual for them to pay or not. Simple as that.

But in my opinion and based on what I’ve seen, I think the ceiling of this Anytwo is just mind-boggling.:star_struck: But, I know right now that I can not fully migrate to it until we have feature parity to Anytype for some of my workflows. But I’m really excited for what’s ahead and I think you guys would be too after you see the Town Hall.

Once again, thanks to the team for their amazing hard work and effort, and I hope the future will be bright for the Any Family!

P.S. I just hope that Anytwo won’t the final name of the product :sweat_smile: but rather an interim internal codename.

7

I’m also been blown away by the bright future that opens up with the new product platform, in terms of development speed, feature richness, and business attractivity (=> well-paying business customers to pay the Team’s bills :slightly_smiling_face: ) — and how much of this bright future has already become visible in the live demos that the Team gave during the Town Hall.

And I can’t stop to congratulate the team taking the courage to rebuild the thing from the ground up —and in ways even closer to the original vision that got me hooked as a beta tester three years ago. I’m so impressed how this courage seems to pay off now, if I only look at the development speed at which this mind-blowing gazillion of long-awaited features (bi-directional block linking, transclusion, rollups, embedded local AI agents, … ) had been brought to live now with a smaller team.

Anytwo finally seems to be the platform now for people to safely keep all of their information and work with it, from PKM users and SMEs to corporations — for everyone who wants to get away from intransparent cloud-based giants like Notion — and now even with fully embedded, sandboxed local AI support in such as smart way (love the “restricted Python” approach). This will be worth a lot to a lot more people — what’s needed now is publicity. Let’s all contribute and spread the word!

Can’t wait to see and test it on my Mac.

8

So if I understand correctly, Anytwo won’t have a “business” version developed for now? Will Antype for business be based exclusively on Anytype?

9

I don’t really understand this. Why wouldn’t the business version be built on the new, more reliable platform instead of the legacy product?

10

Reading the excellent start post, it gives me the same enthusiastic vibe that I’ve felt during the Town Hall! :heart_eyes:

Btw.: Everytime I read the new name “AnyTwo”, I automatically think that it should better be called “AnySquared”.
Because, the new product gives me the feeling that it’s like Anytype “squared”.

There is a whole dimension more on top of nearly each feature that we know.
What previously looked like a one-dimensional list of features, suddenly looks like a two-dimensional area.
It’s not only “some more features” added to the list, but the known features become more depth.

Or, in an other attempt to describe it: it’s like the difference between looking on an image, vs. looking on a hologram.
Or the difference between driving on the given streets (that are full of roadworks and dead ends) vs. flying in a helicopter and enjoying a whole new level of freedom.

“AnySquared”.

11

Thank you for sharing! Exciting changes a looks like it was lots of work. Having watched what I could, here’s some feedback.

General

  • Building on components that exist and have shown to work (like Lexical) well makes full sense. Sometimes, when using Anytype, we wondered why that wasn’t done more often.
  • Personally, I understand the need to progress past what Anytype is currently; perpetually stuck in a beta. Sometimes you just need to reset. Recent developments changed many assumptions about software and the world is moving faster and faster. But - I hope we don’t lose what was worked out over time.
  • It’s good you picked up some features from other tools, like synced blocks. Microsoft Loop ran a cool experiment on this and it worked very well for what it was.
  • Dual graph + tree structure is a game changer. Sign me up now
  • Interoperability and integration will be key. You’re competing against highly integrated/integrable products. AI can be the glue, but won’t be always, and won’t work for everyone.

Editor 2

  • Awesome to see it revamped. This was one of the biggest weaknesses vs. expectations and the competition.

Collections 2

  • Need to test :slight_smile: Grouping looks very helpful.

Mini-apps / self-improvement

  • Self-building and configurable tools are all the rage (PI, BB App…). The right direction IMO.

Specific questions and comments

  1. How would we go about adding an ability to interpret and show geo-coordinates? (e.g. pages/objects on a map, etc). Is this mini app territory? Would this be a core?
  2. Web-based editing without sacrificing e2ee?
  3. Editor comment: It’d be good to have easy creation for tables inline (like just with a Tab key - a la onenote)
  4. Editor comment: It’d be good to have macros - custom blocks that can just be pasted on-demand with a keyboard shortcut or selected from a menu. Alternatively, apply a custom style to a block with a shortcut.
12

I’m overjoyed and excited to see everything from the town hall. A lot of work has been hidden, cooking behind the scenes; and let me tell you, I’m hungry!

Not sure if it was covered, but I can’t wait to see what the designers come up with for the new app icons! As much as I enjoy the current one, I like to spice it up with something new every once in a while.

I’m also curious to see the direction this community takes. Will it become solely dedicated to Anytwo, and Anytype conversations will move to other platforms like GitHub?

13

I have a quick architectural question: does Anytwo move away from Electron, or is it still built on it?

Personally, I’m really not a fan of Electron. To me, it carries its own legacy baggage and could easily become a bottleneck for performance and future development. Given that you’re rebuilding the foundations from the ground up, did that clean-slate approach include ditching the Electron architecture as well?

14

The first test build isn’t out yet and already he’s making new feature requests :joy::joy::joy::joy:

15

Satoru Iwata Former CEO of Nintendo spent years developing a game called Earthbound. Due to the messy code Satoru decided to rewrite the entire code 6 months from launch and now its one of the most popular games from the 90s.

I see Anytwo as no different and looking forward to what you’ll come up with!

Can we get Anytwo before GTA 6? :winking_face_with_tongue:

16

This is exactly what I feared.
So once again, we have a product in its very early stages, and the whole process of feature requests and bug reports starts all over again. The promise that everything will progress very quickly this time, isn’t really reassuring. Anytype didn’t even make it to version 1.0, and now we’re starting with Anytwo at an even earlier stage. With the added side effect that all existing data — which was painstakingly entered — will have to be migrated, and we don’t know when this will be possible without risk.

“We plan to release the product in its first iteration publicly by the end of the year.” - Experienced users who are familiar with following software development know that this could just as easily be the middle of next year if, for example, unexpected difficulties arise.

Anytwo, which will surely shine at some point in the future, is for people who have a great deal of time, patience, love, and hope.

In my opinion, when it comes to making the product profitable and acquiring new, paying users, the decision to develop a completely new product with a new codebase should have been made much, much earlier.
As it stands now, the problem is bigger than it’s ever been. Either keep riding a dead horse or to embark on the arduous journey of seeing the entire development process through to a truly usable product — neither is a tempting choice.

Still, I’m looking forward to testing Anytwo.

17

Super excited for this and the potential future direction of the project! Hopefully I can get in on the tests, I think it’s fun to be a part of things like that!

18

Will the software we developed for Anytype Local API work on a new product?

19

I have just finished watching the Town Hall replay and, like many, I oscillated between apprehension and excitement. The demo really got me excited, and the new editor, my major actual pain point, looks absolutely fantastic.

However, having lived through the period when Anytype was still in alpha, then the first migration to the current version, as well as all the iterations, the idea of starting again from scratch is quite scary to me as a user who relies on this tool every day, which has finally been quite stable.

In short, like many, I look forward to getting my hands on the software to get an proper idea.

20

As much as I’m excited, I also fully share @VisualNotes 's point of view.
The release of the first iteration at the end of the year sounds already hard. But OK, that’s as it is, good things need time.
BUT: that there will no feature parity in the first iteration, sounds REALLY bad!

I was already about to write about that, thanks to VisualNotes that he also addresses to exactly this point.
Honestly, even if the first release takes more time, it should definitely have all of Anytype’s features from the first iteration on!
Otherwise we would end up with two half-baked products in parallel, where a migration doesn’t make sense and we don’t know which version we should use to put our new Objects in.

In my opinion, feature parity from start on should have the highest priority over fancy new features, so that we are able to migrate as early as possible.
It that is given, and we all have successful migrated, then can accept to wait maybe some months more for the additional fancy new features.