I’m delighted to see this project evolving and look forward to following its development! I’ve used just about every markup technology over my (many) years as a tech writer, from troff macros, to SGML/DocBook and then XML/DITA, and finally, to Markup with the Material for MkDocs project, as a sponsor. Each has its strengths and weaknesses, but for enabling contributors outside the tech writing community, simpler source formats are the way.
That said, if pressed, I’d recommend AsciiDoc[0] over any Markup flavor for a greenfield project _today_. We had to either add or bake plugins or extensions to get features that are already included in AsciiDoc, making our Markup implementation both more complex and wholly unique. That wasn’t a huge problem, because we didn’t have a large pool of contributors to educate and support, but it would have been much easier just to point to a standard.
But, hey! The roadmap includes modules, to make way for other source formats! This is the way. :-)
> That said, if pressed, I’d recommend AsciiDoc over any Markup flavor for a greenfield project _today_.
Likewise for me as well, and I am a massive Material for MkDocs fan. Markdown is certainly simple to use and gets the job done, but AsciiDoc just provides so much out of the box without hurting my eyes like reStructuredText (used by Sphinx) does. It also helps that's there's effectively one type of AsciiDoc I'm aware of, whereas there's a number of Markdown flavors atop CommonMark to be cognizant of. I will concede, however, it's learning curve is not as simple as MarkDown's...
A powerful framework for working with AsciiDoc for documentation purposes is Antora[0]. The Red Hat ecosystem (Fedora and CentOS projects) uses it for their public facing docs. That being said, it is a beast to understand if starting from scratch rather than contributing to project's existing docs. It designed to be able to consolidate large projects with multiple component repositories and versions per component into a single docs site. Typical balance of more capabilities, more up-front cost of adoption.
The AsciiDoc WG also maintains an Awesome AsciiDoc[1] page of projects within the ecosystem.
It is a very reasonable choice: your plug-in project has grown to a scale you have never anticipated. Therefore, you built your own system from the ground up, one which you have total control over, instead of keeping to rely on someone else's system as you did before. This would make new features and bug fixes more easy to be implemented.
I am also very curious about what the MKDocs future would be like. Material has been the most popular theme for MkDocs. People are not using it because they have chosen MkDocs, but using MkDocs because they have chosen Material. With Zensical promising (some kind of) MkDocs compatibility, there will be fewer reasons to stay on MkDocs instead of migrating to the new project.
> It is a very reasonable choice: your plug-in project has grown to a scale you have never anticipated. Therefore, you built your own system from the ground up, one which you have total control over, instead of keeping to rely on someone else's system as you did before. This would make new features and bug fixes more easy to be implemented.
Exactly this. We ran against walls trying to realize our ideas with MkDocs' APIs, so a rewrite was due. With MkDocs being unmaintained for over a year, we took the initiative. Since we have excellent product-market fit, investing into a new stack was just logical.
This looks very interesting, I'm using mkdocs+material for one site (and it's great) but trying to find a good solution for more complex docs.
Is there anything planned like 11ty's data files[1]? For example, I can pass it a JSON file, or a TOML file, and have it generate one HTML page per document in that file using their pagination system in a hacky way, and add those pages to collections for grouping in navigation and such using their templating language. This makes it a lot easier to autogenerate documentation from a combination of sources (since anything we want can emit some appropriate JSON/TOML) without having a real "source document". It's hard to tell at a glance if this is something that would be possible with the upcoming module system.
I'm really excited by this development! Material for MkDocs has raised the quality level of so many projects' docs, my own included, by making good navigation the default. It's by far my favorite system to browse as a reader, and use as a project maintainer.
I hope the new theme allows for more customization than the old Material theme. It was really hard to create a unique brand identity within the constraints of Material; it just wasn't built with customization in mind beyond a color. The "modern" theme looks minimal in a way that gives me some hope for this.
I'm really intrigued by the use of differential dataflow in a static site toolkit, but there isn't much in the way written about it. If anyone from the team are here I would love it if you could explain how it's being used? Does this enable fast incremental builds, only changing the parts that change in the input? If so how do you model that? Are you using multisets as a message format inside the engine?
For context, I work on TanStack DB which is a differential dataflow / DBSP inspired reactive client datastore. Super interested in your use case.
Excellent question. We're not using differential dataflow (DD), but are rolling our own differential runtime. It's basically functions stitched together with operators, heavily inspired by DD and RxJS, and is optimized for performance and ease of use. The decision to go from scratch allows us to provide something that, IMHO, is much simpler to work with than DD and Rx, as our goal is to allow an ecosystem to evolve, as MkDocs did as well. For this, the API needs to be as simple as possible, while ensuring DD semantics.
Right now, builds are not fast, since Python Markdown is our bottleneck. We decided to go this route to offer the best compatibility possible with the Material for MkDocs ecosystem, so users can switch easily. In the next 12 months, we'll be working on moving the rest of the code base gradually to Rust and entirely detaching from Python, which will make builds significantly faster. Rebuilds are already fast, due to the (still preliminary) caching we employ.
The differential runtime is called ZRX[1], which forms the foundation of Zensical.
I was excited up until they showed what the new theme looks like. mkdocs-Material was nice in that it didn't have overly rounded corners and over travesties, a shame that custom CSS will be needed to undo the "modernisation". Overall this seems very interesting, especially the performance improvements, just a letdown visually.
Creator of Zensical here! As always, it's a matter of taste. We felt the original look was a little date. You can use the classic Material for MkDocs look with Zensical by adding a single line of configuration[1]. This works because the HTML is exactly the same right now. Most users favor the new look over the old one.
From the feedback we got after launching. More accurately: most users that we conversed with since launch are very happy about the new look. Regardless, for compatibility reasons, the old look is available as well.
Someone builds a complete documentation system from scratch and you call it a "travesty" because it has rounded corners? Visual design is important but this is very misplaced priorities.
In todays News, someone built a whole Hotel with freely available design plans, and a user finds a travesty in the font style of the Cold/Hot sink taps being too round.
I really like this new approach and I will give it a go on some open-source projects I have. Btw, I have looked into Zensical Spark but I did not fully understand what is it? It is just for connecting with the team so you help other teams to set it up and help them with feature requests and training?
Zensical Spark is our offering for professional users. In the past 12 months, we've had numerous conversations with organizations and enterprises to understand how they adapted Material for MkDocs to their processes and workflows. Our goal with Zensical is to build a far more flexible solution than (Material for) MkDocs, developed closely alongside the professional users who rely on it daily. This is a paid product and represents the way we ensure the long-term sustainability of the project.
Can I also use it to write blog posts and generate an RSS feed for them?
A lot of documentation sites want this but it was always hard with things like sphinx where the input must be files on disk so one couldn't e.g. load blog posts from a db.
Thanks for the information. If files on disk are a requirement that would be a shame. Such tools keep abstracting such that one can generate a plethora of output formats but don't abstract the input in such a way that one could programmatically add content. Even though that wouldn't stand in the way of using the filesystem as the source regardless.
Unfortunately, I require PDF outputs for some of my documentation. Right now this is possible via some plugins (I cannot remember which fork of a pdf export plugin worked, but at least one does) in Material for MkDocs. It’s not perfect, but it is good enough.
Should I expect a “good enough” pdf export experience in zensical at some point or now?
Yes, we're basically agnostic to the input and output formats. Right now it's Markdown -> HTML, but with the upcoming module system, it'll be possible to convert anything to anything. Our focus will stay Markdown/HTML first, and once we reach feature parity, we'll explore to support formats like PDF etc. natively.
the featured search engine is not typo-resistant, so unfortunately loses to google again, a bane of many sites. Wish search baseline finally caught up to the fact that humans can't type perfectly... (hopefully that's just because it's new and fresh, though)
It's definitely not malicious intent. It's an inlined version of our new search engine that we'll release in early 2026, but already wanted to ship with Zensical. However, you're right that this might raise some eyebrows – we'll fix it with the next release.
I use docusaurus and am happy - any reason I should switch to this or does it make more sense if you have a greenfield use case?
I don’t really care how long it takes to build as long as it’s not minutes.
Also don’t care about “modern” look whatever that means really.
And I have lots of custom components that (I assume?) would be hard to migrate
Material Design have been overhauled twice since the theme has been released. That was not the reason why they succeeded, just a solid foundation which they happened to built from.
Documentation tooling for Python (especially reference docs) is… abysmal. Sphinx is the king, with no markdown support. Pdoc would be nice, it the authors didn’t reject reasonable additions, leading to forks. And mkdocs is not simple to setup. If you’re building a company for this, please fix this.
AFAIK, Sphinx does support markdown… the issue with markdown is that it doesn’t support macros like RST, so there’s a huge amount of functionality that can’t translate well to it.
We'll do our best! The author of mkdocstrings[1] (the leading API reference docs solution in the MkDocs space) is part of our team, so we're aiming to bring much more powerful API reference docs support to Zensical.
The biggest issue I had with material for mkdocs was training BA type people to contribute. They had to muddle their way through python, pip, git, github, etc. just to make a one line change.
We've heard this many times when talking to our enterprise users, which is one of the reasons that motivated the fresh start. WYSIWYG is on our roadmap as a stretch goal, allowing non-tech users to contribute. It'll take some time, but we'll reach it eventually!
i actually tried zensical yesterday and fell back to mkdocs, not sure what I'm missing while trying. Material for mkdocs served me well so I will stay with it for now.
zensical installation is longer than mkdocs, probably due to the rust side.
I’m delighted to see this project evolving and look forward to following its development! I’ve used just about every markup technology over my (many) years as a tech writer, from troff macros, to SGML/DocBook and then XML/DITA, and finally, to Markup with the Material for MkDocs project, as a sponsor. Each has its strengths and weaknesses, but for enabling contributors outside the tech writing community, simpler source formats are the way.
That said, if pressed, I’d recommend AsciiDoc[0] over any Markup flavor for a greenfield project _today_. We had to either add or bake plugins or extensions to get features that are already included in AsciiDoc, making our Markup implementation both more complex and wholly unique. That wasn’t a huge problem, because we didn’t have a large pool of contributors to educate and support, but it would have been much easier just to point to a standard.
But, hey! The roadmap includes modules, to make way for other source formats! This is the way. :-)
[0] https://asciidoc.org/
EDIT: s/That's the way/This is the way. :-)/
> That said, if pressed, I’d recommend AsciiDoc over any Markup flavor for a greenfield project _today_.
Likewise for me as well, and I am a massive Material for MkDocs fan. Markdown is certainly simple to use and gets the job done, but AsciiDoc just provides so much out of the box without hurting my eyes like reStructuredText (used by Sphinx) does. It also helps that's there's effectively one type of AsciiDoc I'm aware of, whereas there's a number of Markdown flavors atop CommonMark to be cognizant of. I will concede, however, it's learning curve is not as simple as MarkDown's...
A powerful framework for working with AsciiDoc for documentation purposes is Antora[0]. The Red Hat ecosystem (Fedora and CentOS projects) uses it for their public facing docs. That being said, it is a beast to understand if starting from scratch rather than contributing to project's existing docs. It designed to be able to consolidate large projects with multiple component repositories and versions per component into a single docs site. Typical balance of more capabilities, more up-front cost of adoption.
The AsciiDoc WG also maintains an Awesome AsciiDoc[1] page of projects within the ecosystem.
[0] https://antora.org/ [1] https://gitlab.eclipse.org/eclipse-wg/asciidoc-wg/asciidoc.o...
It is a very reasonable choice: your plug-in project has grown to a scale you have never anticipated. Therefore, you built your own system from the ground up, one which you have total control over, instead of keeping to rely on someone else's system as you did before. This would make new features and bug fixes more easy to be implemented.
I am also very curious about what the MKDocs future would be like. Material has been the most popular theme for MkDocs. People are not using it because they have chosen MkDocs, but using MkDocs because they have chosen Material. With Zensical promising (some kind of) MkDocs compatibility, there will be fewer reasons to stay on MkDocs instead of migrating to the new project.
> It is a very reasonable choice: your plug-in project has grown to a scale you have never anticipated. Therefore, you built your own system from the ground up, one which you have total control over, instead of keeping to rely on someone else's system as you did before. This would make new features and bug fixes more easy to be implemented.
Exactly this. We ran against walls trying to realize our ideas with MkDocs' APIs, so a rewrite was due. With MkDocs being unmaintained for over a year, we took the initiative. Since we have excellent product-market fit, investing into a new stack was just logical.
As a multi-year sponsor for your work, our open-source community really appreciates the work.
This looks very interesting, I'm using mkdocs+material for one site (and it's great) but trying to find a good solution for more complex docs.
Is there anything planned like 11ty's data files[1]? For example, I can pass it a JSON file, or a TOML file, and have it generate one HTML page per document in that file using their pagination system in a hacky way, and add those pages to collections for grouping in navigation and such using their templating language. This makes it a lot easier to autogenerate documentation from a combination of sources (since anything we want can emit some appropriate JSON/TOML) without having a real "source document". It's hard to tell at a glance if this is something that would be possible with the upcoming module system.
[1]: https://www.11ty.dev/docs/data-global/
I'm really excited by this development! Material for MkDocs has raised the quality level of so many projects' docs, my own included, by making good navigation the default. It's by far my favorite system to browse as a reader, and use as a project maintainer.
I hope the new theme allows for more customization than the old Material theme. It was really hard to create a unique brand identity within the constraints of Material; it just wasn't built with customization in mind beyond a color. The "modern" theme looks minimal in a way that gives me some hope for this.
Looking forward to kicking the tires on Zensical!
I'm really intrigued by the use of differential dataflow in a static site toolkit, but there isn't much in the way written about it. If anyone from the team are here I would love it if you could explain how it's being used? Does this enable fast incremental builds, only changing the parts that change in the input? If so how do you model that? Are you using multisets as a message format inside the engine?
For context, I work on TanStack DB which is a differential dataflow / DBSP inspired reactive client datastore. Super interested in your use case.
Excellent question. We're not using differential dataflow (DD), but are rolling our own differential runtime. It's basically functions stitched together with operators, heavily inspired by DD and RxJS, and is optimized for performance and ease of use. The decision to go from scratch allows us to provide something that, IMHO, is much simpler to work with than DD and Rx, as our goal is to allow an ecosystem to evolve, as MkDocs did as well. For this, the API needs to be as simple as possible, while ensuring DD semantics.
Right now, builds are not fast, since Python Markdown is our bottleneck. We decided to go this route to offer the best compatibility possible with the Material for MkDocs ecosystem, so users can switch easily. In the next 12 months, we'll be working on moving the rest of the code base gradually to Rust and entirely detaching from Python, which will make builds significantly faster. Rebuilds are already fast, due to the (still preliminary) caching we employ.
The differential runtime is called ZRX[1], which forms the foundation of Zensical.
[1]: https://github.com/zensical/zrx
I was excited up until they showed what the new theme looks like. mkdocs-Material was nice in that it didn't have overly rounded corners and over travesties, a shame that custom CSS will be needed to undo the "modernisation". Overall this seems very interesting, especially the performance improvements, just a letdown visually.
Creator of Zensical here! As always, it's a matter of taste. We felt the original look was a little date. You can use the classic Material for MkDocs look with Zensical by adding a single line of configuration[1]. This works because the HTML is exactly the same right now. Most users favor the new look over the old one.
[1]: https://zensical.org/docs/setup/basics/#theme-variant
> Most users favor the new look over the old one.
How do you know?
From the feedback we got after launching. More accurately: most users that we conversed with since launch are very happy about the new look. Regardless, for compatibility reasons, the old look is available as well.
Early adopters are a very distinct set of users. Probably not great to extrapolate from it.
Someone builds a complete documentation system from scratch and you call it a "travesty" because it has rounded corners? Visual design is important but this is very misplaced priorities.
In todays News, someone built a whole Hotel with freely available design plans, and a user finds a travesty in the font style of the Cold/Hot sink taps being too round.
I really like this new approach and I will give it a go on some open-source projects I have. Btw, I have looked into Zensical Spark but I did not fully understand what is it? It is just for connecting with the team so you help other teams to set it up and help them with feature requests and training?
Zensical Spark is our offering for professional users. In the past 12 months, we've had numerous conversations with organizations and enterprises to understand how they adapted Material for MkDocs to their processes and workflows. Our goal with Zensical is to build a far more flexible solution than (Material for) MkDocs, developed closely alongside the professional users who rely on it daily. This is a paid product and represents the way we ensure the long-term sustainability of the project.
Zensical itself is entirely free OSS software.
Can I also use it to write blog posts and generate an RSS feed for them?
A lot of documentation sites want this but it was always hard with things like sphinx where the input must be files on disk so one couldn't e.g. load blog posts from a db.
Its predecessor (mkdocs material) supported blogs via a plugin and rss feeds via a mkdocs plugin (https://guts.github.io/mkdocs-rss-plugin/). I don't believe it supports blogs yet according to this: https://zensical.org/compatibility/#phased-transition-strate... but will in the next phase when it supports plugins.
It doesn't support loading content from a database that I'm aware of, it uses markdown files on disk as input.
Thanks for the information. If files on disk are a requirement that would be a shame. Such tools keep abstracting such that one can generate a plethora of output formats but don't abstract the input in such a way that one could programmatically add content. Even though that wouldn't stand in the way of using the filesystem as the source regardless.
Interested to look into this more, but Astro Starlight[1] has been by far the best in recent times. Only Vitepress comes close.
[1]: https://starlight.astro.build/getting-started/
The new modern static site generator and best for now seems to be Jupyter Book 2 - https://jupyterbook.org/stable/
Launched last week and build upon MyST engine. Makes Shphinx obsolete. Like to see a real comparison with Zensical.
Unfortunately, I require PDF outputs for some of my documentation. Right now this is possible via some plugins (I cannot remember which fork of a pdf export plugin worked, but at least one does) in Material for MkDocs. It’s not perfect, but it is good enough.
Should I expect a “good enough” pdf export experience in zensical at some point or now?
Yes, we're basically agnostic to the input and output formats. Right now it's Markdown -> HTML, but with the upcoming module system, it'll be possible to convert anything to anything. Our focus will stay Markdown/HTML first, and once we reach feature parity, we'll explore to support formats like PDF etc. natively.
https://github.com/zensical/zensical
the featured search engine is not typo-resistant, so unfortunately loses to google again, a bane of many sites. Wish search baseline finally caught up to the fact that humans can't type perfectly... (hopefully that's just because it's new and fresh, though)
We'll ship fuzzy search in the coming weeks. It's just awaiting a release.
Nice!
Hi, congrats on Zensical release. Want to ask is there any community space where people discuss ideas or what's being built?
if not I'd be happy to help get something started for folks who want to learn and contribute.
We set up a Discord for the community, and have a dedicated space for professionals! https://zensical.org/docs/community/get-involved/
Why are some files obfuscated in the ui repo? It looks like compiled typescript has been included without the original source.
A gated feature? Malicious intent?
It's definitely not malicious intent. It's an inlined version of our new search engine that we'll release in early 2026, but already wanted to ship with Zensical. However, you're right that this might raise some eyebrows – we'll fix it with the next release.
I use docusaurus and am happy - any reason I should switch to this or does it make more sense if you have a greenfield use case?
I don’t really care how long it takes to build as long as it’s not minutes. Also don’t care about “modern” look whatever that means really. And I have lots of custom components that (I assume?) would be hard to migrate
Bold to move away from both of the things that contributed to their success initially (mkdocs, material design) at once.
Material Design have been overhauled twice since the theme has been released. That was not the reason why they succeeded, just a solid foundation which they happened to built from.
Documentation tooling for Python (especially reference docs) is… abysmal. Sphinx is the king, with no markdown support. Pdoc would be nice, it the authors didn’t reject reasonable additions, leading to forks. And mkdocs is not simple to setup. If you’re building a company for this, please fix this.
AFAIK, Sphinx does support markdown… the issue with markdown is that it doesn’t support macros like RST, so there’s a huge amount of functionality that can’t translate well to it.
We'll do our best! The author of mkdocstrings[1] (the leading API reference docs solution in the MkDocs space) is part of our team, so we're aiming to bring much more powerful API reference docs support to Zensical.
[1]: mkdocstrings.github.io
Sphinx does have markdown support fwiw
It looks interesting, might give it a try. I currently use Zola and it’s been great so far.
The biggest issue I had with material for mkdocs was training BA type people to contribute. They had to muddle their way through python, pip, git, github, etc. just to make a one line change.
We've heard this many times when talking to our enterprise users, which is one of the reasons that motivated the fresh start. WYSIWYG is on our roadmap as a stretch goal, allowing non-tech users to contribute. It'll take some time, but we'll reach it eventually!
now adding rust on top of that.
i actually tried zensical yesterday and fell back to mkdocs, not sure what I'm missing while trying. Material for mkdocs served me well so I will stay with it for now.
zensical installation is longer than mkdocs, probably due to the rust side.