About

MatGen is a dialectical knowledge management and content creation system.

By leveraging a clean, semantic database structure, it allows users to keep all their data in one place, track any change semantically and automatically, and fully integrate with different functionality, such as project management, website generation, automatic typesetting and much more.

It uses a unique approach to data management, by using folders as semantic entities, and forgoing relational databases in favour of using raw files directly.
This virtually eliminates data duplication, removes technical friction for users, ensures interoperability with any and all existing software, allows for the dynamic reuse of data across modules, and makes automatic semantic versioning, backups, debugging and documentation a reality.
Various system optimisations ensure high performance and fast dynamic indexing.
Standardised, validating ontologies guarantee data cleanliness and AI-ready semantics, as well as inherent accessibility.

The semantic core is the perfect base for innumerable modules, and is an important step towards building a truly semantic and accessible web, where knowledge is built collectively and shared.

What fundamentally distinguishes MatGen from other software is its core system and the reach of its functionality.
It is a self-maintaining development ecosystem where adding functionality automatically generates ad-hoc documentation, updates semantic relationships, and preserves system-wide consistency. Functionality is constellational, not hierarchical, and this allows the easy building of cross-domain modules.
No other software is built specifically to handle everything: it can be used as a CMS, with integrated CRM and project/note management, outputting accessible websites, typeset books, title cards for videos, print-ready promotional material, ensemble-specific sheet music with player-specific information packs for performances, anything you can imagine.

MatGen's folder-per-entity structure takes the directness and efficiency of SSGs, while adding rich ontologies and database operations to the raw filesystem, essentially providing a rich content management system that can virtually take any input and produce any output.
This retains the ability to use GIMP, Inkscape, Krita, Blender, MuseScore or any software to work on files, while also having automatic semantic versioning and renders.
This means that anyone can build interfaces to handle MatGen data, and due to the AGPLv3 licensing vendor lock-in is legally impossible.

Its universal granular theming system takes inspiration from KDE Plasma's theming system and modern CSS variable-based systems, but adds even more granularity. Theming is semantics-based, not class or ID based.

Templates in MatGen are not prescriptive and end-to-end like LaTeX, Typst, Hugo, but pre-built semantic structures (essentially pre-built folders) which can be altered granularly thanks to the core component system, which is language-agnostic, much like all MatGen data. This means that users can create one UI structure, and that it can be deployed across languages and platforms, it means that the same exact structure or layout can be used across themes, projects and entities.
This means that MatGen data is future-proof, and that all MatGen data will remain compatible, even as the logic code evolves, not too dissimilar to rpm-ostree's philosophy.

Project management and note-taking happens in your files, not in a separate database. Tasks are semantic folders with deep metadata. This opens up literally infinite scripting possibilities. Notes live in the folders they refer to, unlike LogSeq, Joplin, OneNote and the likes. There is no need for vaults like in Obsidian. Seeing as every folder is a semantic entity, it's trivial to show notes per entity in the GUI, and have various views based on various metadata and filters to find cross-domain notes, as well as semantic metadata and ontologies: MatGen brings MVC to note management.

MatGen's semantics are the reason it can do all this. Total separation of concerns means that data can be created, edited and viewed in infinite ways, depending on specific needs.
This approach also means inherent accessibility, as there can be no dirty data, by design.
This also means that outputs are deterministic, unlike AI-based solutions. There can be no hallucinations, and future AI-systems will be able to benefit from clean, structured and validated data.

At the root of its philosophy is the recognition that domains are descriptive, not prescriptive. Instead of having separate modules for functionality that is virtually identical except for the domain of the data it handles, MatGen separates modules by the output format, because that's when they require truly different code.
The use of open standards and the focus on parsers means that data is open and available to everyone for any purpose: data is data, and it should be referenceable anywhere, much like in Ted Nelson's vision for Project Xanadu.

MatGen is currently under development, with a projected initial release date somewhere in Q4 of 2026.

For any enquiries, get in touch at enquiries@matgen.app .