This is where the open ecosystem for digital law is taking shape.
With rulemapping, a new era of legislation and legal application begins. Laws are no longer created solely as text, but simultaneously as visual, machine-readable logic models. This allows them to be processed, validated, and applied digitally from the outset. This open ecosystem connects legal scholarship, public administration, and technology — laying the foundation for a digital, transparent, and effective legal system.
Law as Code is the idea of designing law digitally from the start — creating legislation in a way that machines can process and apply directly within digital procedures. Rulemapping is the first market-ready implementation of this concept: it translates laws into visual, legally structured decision models that are both machine-readable and understandable for humans. This establishes the basis for a state that not only becomes digitalized but functions digitally — with laws that update automatically, decisions that remain explainable, and processes that act transparently and consistently.
With the Rulemap Builder (BETA), laws, guidelines, and internal regulations can already be modeled as digital decision structures — free of charge, visually, and without any programming. Further development stages will follow soon: the open Rulemap Library as a shared platform and the integration of Rule AI for automated application. Anyone working with rulemapping today actively shapes the foundation of tomorrow’s open, digital legal system.
The Rulemap Builder (BETA) is the first building block of this ecosystem. It enables laws, regulations, and procedures to be represented and reviewed as visual decision models. It forms the precursor to automating legal processes with Rule AI — a technology that makes legal logic machine-executable. Future development stages will allow AI-assisted generation of Rulemaps, automatic modeling of existing laws, and the creation of legally compliant texts based on the Handbuch der Rechtsförmlichkeit — all compatible with standards such as LegalDocML or FIM.
The Rulemap Library will become an open collection of digital laws and rule models — freely accessible, extendable, and interoperable. It allows Rulemaps to be shared, compared, and further developed. For the first time, this creates a collaborative legal infrastructure that brings machine-readable law to life step by step.
RUML (Rulemapping Logic Format) is an open, JSON-based file format that provides a machine-readable representation of laws, regulations and decision logic. It makes legal rules understandable to software without altering their legal meaning.
What problem does RUML solve?
Currently, laws and regulations are mostly available in plain text format. This results in manual and error-prone interpretation whenever they are applied, whether in public administration, business or software development. RUML overcomes this issue by translating legal rules into a standardised, machine-readable format that can be evaluated automatically.
Who is RUML designed for?
RUML is intended for developers building parsers, validators, editors or other tools for processing legal logic. Beyond that, legislators, public administrations and businesses benefit from modelling their regulations digitally and applying them automatically.
Is RUML freely available?
Yes. The RUML specification is open and will be published for public access, enabling governments, organisations and developers worldwide to use and integrate the format into their systems free of charge.
Where can RUML be used?
Anywhere rule-based decisions are made, such as in tax law, social benefits, building permits, compliance checks, insurance logic or internal corporate policies. RUML is designed to be independent of jurisdiction and industry.
These build on the core functions of the Rulemap Builder, providing modular extensions that simplify and optimise rulemap workflows for specific application scenarios. While not strictly necessary for modelling alone, they enhance the Rulemap Builder with additional functions that speed up, automate or improve the quality of certain work steps. The Builder itself remains fully usable in its basic version. The VAS are premium modules that can be activated as and when required, with pricing based on usage or licence, depending on the module. While most of these modules are AI-powered, they are designed to be model-agnostic, meaning they function independently of the structure or type of machine learning model used.
The underlying language models are interchangeable and can be integrated with existing legacy systems in on-premises setups — a prerequisite for operating the modules in regulated environments without external model dependency. The two core value-added services are Arum and Mura: two AI-powered add-ons in public beta that facilitate two-way translation between regulatory text and logic structure.
Transforming text into Rulemaps
Arum automates the most time-consuming initial step in any modelling process, converting existing regulatory text into an initial rulemap. Rather than manually extracting the decision-making structure from a set of paragraphs, users are provided with an AI-generated logic tree reflecting the structure of the source text. This tree can then be edited directly within the Rulemap Builder. As work progresses, this logic tree is developed into the actual decision tree. Arum provides the starting point, but not the finished model. Arum works most reliably with fully formulated legal texts. Although policy papers, guidelines, and other regulatory frameworks can also be processed, the accuracy of the result decreases as the source structure becomes more open-ended. The maximum input length is currently 10,000 characters.
The module is in a public beta phase and can be used free of charge during this period. Pricing at a later date cannot be ruled out. Free use is limited to a certain number of runs per user. This quota can be increased on request. The rollout is taking place in stages, so not all users will have access from the outset. The structures generated by Arum are not intended to be accurate or complete. They serve as a starting point for further processing in the Rulemap Builder, whilst the business-specific refinement into a validation structure takes place in the subsequent modelling steps.
From Rulemap to text
While Arum streamlines the creation of rulemaps, Mura works in the opposite direction. The module generates natural-language text from an existing rulemap, whose content and logic adhere to the underlying structure. Mura therefore provides users with the editorial foundation on which they can draft legislation, consolidated versions of regulations or accompanying explanatory texts. This addresses a particularly time-consuming step in the legislative process: translating a logical structure into consistent legal language. Mura produces the initial editorial draft, but the user is responsible for the final textual form.
Like Arum, Mura is currently in a public beta phase, offering free use with a limited usage quota that can be increased on request. Pricing may be introduced at a later date, with access rolled out in stages. The results are drafts, not final products. They do not replace legal or editorial review and make no claim to accuracy. Mura accelerates the transition from logic to language, but users remain responsible for the final text.
