News / public record

Checking
  1. Update / Component / Status · CheckingLegal Studio: technical profile confirmedLegal Studio 0.1.0 is in active development, preparing for the MVP. Its code remains private, with the same individually agreed partnership terms as Legal Workspace.
  2. Update / Component / Scope · CheckingOLC Engine: a renewed presentation of legal workThe bilingual Engine page now presents the approved version 0.1.0 scope, its legal-work method and the boundary between current capabilities and planned expansion.
  3. Article · CheckingA Server in the Office Is Not a Security StrategyWhy an office server, cloud service, email route or encrypted backup must be assessed through capabilities, responsibility and evidence—not location alone.
  4. Article · CheckingWho Controls the Legal AI Stack?Public code can enable inspection without proving control of the whole system. Follow versions, provenance, data sovereignty, exit and regulatory roles in legal AI.
  5. Article · CheckingConfidentiality Does Not End at the Chat WindowConfidentiality is a property of the complete data path, not the chat window. What do DPAs, de-identification and local deployment solve — and why is the OLC gateway still only a concept?
  6. Article · CheckingWhen a High Score Hides a Critical Legal ErrorA high benchmark score does not establish that legal AI is dependable for every problem. This article explains how to measure critical errors, retrieval, abstention, evaluators and the path to a legal answer.
  7. Update / Component / Readiness · CheckingWord Connector public presentation is completeThe Apache-2.0 v0.1.0-beta.1 source-only release now has a complete public component record, homepage feature, Roadmap outcome and discovery path.
  8. Update / Component / Release · CheckingOpenLegalCore Word Connector v0.1.0-beta.1 is publicThe Apache-2.0 source-only beta provides a verified Word for the web path for bounded research into Slovenian legal sources through OLC Engine, returning source-linked answers without changing the document.
  9. Article · CheckingLegal RAG Can Fail Before the Model AnswersA genuine provision can still support the wrong answer. This article shows how collection scope, document splitting, filters and ranking determine which sources a legal AI model ever sees.
  10. Article · CheckingWhen a Correct Citation Leads to the Wrong LawA genuine provision and an official link do not prove that the right law was applied. This article explains why legal AI must preserve temporal versions, transitional rules and the source-selection path.
  11. Update / Project · CheckingPublic project foundation gate is completeOpenLegalCore has completed its public project foundation gate: core public routes, bounded participation channels and publication controls are operating.
  12. Article · CheckingAn Audit Trail Is Not a Log: What a Digital Legal Process Must PreserveAn audit trail is useful when it permits a reasonable reconstruction of the particular process while remaining purpose-bound, appropriately protected and subject to retention rules.
  13. Article · CheckingArticle 86 of the AI Act: Does the right to explanation already apply to Annex III systems?The legal question remains open. The technical capacity to provide an intelligible explanation should not.
  14. Update / Project · CheckingPrivate security reporting is now availableOpenLegalCore now provides a monitored project-wide private channel and a canonical security-reporting policy.
  15. Update / Component / Release · CheckingSlovenian Case Law Pipeline v0.1.7 is publicOpenLegalCore has published the production-verified Slovenian case-law ingestion component as a source-available BUSL-1.1 release.
  16. Article · CheckingA Result Is Not a Method: Why Visible Methods Matter in Legal AIWhy legal AI needs inspectable sources, provenance and human review: a legal and technical analysis of the EU Artificial Intelligence Act and GDPR.
  17. Update / Component / Release · CheckingSlovenian Legislation Pipeline v0.1.0 is publicOpenLegalCore has published the production-verified PISRS legislation-ingest component as a source-available BUSL-1.1 release.
  18. Update / Component / Release · CheckingLegal OCR Pipeline v0.1.2 is publicThe first public OpenLegalCore component is available with code, tests, offline review tooling and a bounded acceptance record.
Let’s talk
Menu
Let’s talk

Page map

01 / 09

OpenLegalCore / legal-tech infrastructure

Open-source components. A private, reviewable system core.

The law is too important for black-box systems.

Inspect the code. Control the deployment. Verify the method.

OpenLegalCore is developing a legal AI engine tailored to the law of a given jurisdiction, along with supporting components for legal analysis and drafting. It brings GOVERNMENT WEBSITE / GOVERNMENT OFFICE FOR LEGISLATIONPISRS — Legal Information System of the Republic of SloveniaA government website providing access to legislation, case law and other legal information from Slovenia and the EU. It is administered by the Government Office for Legislation. Unofficial consolidated texts make legislation easier to consult, but are not official publications. PISRS also provides access to authentic electronic editions of the Official Gazette of the Republic of Slovenia from 1 March 2026.Open PISRS , OFFICIAL SOURCE / SUPREME COURT OF THE REPUBLIC OF SLOVENIASlovenian case lawSlovenia's authoritative national public case-law collection, published through the judiciary's official service administered by the Supreme Court of the Republic of Slovenia. It is the complete public corpus available through that service; statutory publication rules do not require every decision issued by every Slovenian court to appear.Open the official collection and private documents together in controlled, traceable workflows whose operation and results remain open to human review.

Explore the OLC system· Discover its parts interactively

The central partially open OLC Engine shows four linked functional layers: legal foundation and sources, case context and material, analysis and strategy, and drafting with traceable outputs. Professional review remains human-led. The Engine links legal sources, document processing and work environments.

An illustration of the system's composition, not a depiction of actual infrastructure.

Who OpenLegalCore is for

The same legal method.Different ways to use it.

Legal work takes place in different environments. A law firm may need an organised workspace for matters and documents. An insurer, bank or public institution wants to integrate the legal method into a system it already uses. OpenLegalCore therefore separates the legal core, workspace, jurisdiction-specific foundation and integrations. We connect them according to the workflow, deployment approach and rules for handling data.

01Legal workLegal Workspace · Active development · preparing the MVP

A workspace for legal professionals

Bring conversations, matters, documents and notes into one clear workspace. Legal Workspace is designed for independent lawyers, legal teams and law firms. It is being developed to support shared and private spaces, helping you review findings, drafts and decisions as the work progresses.

Explore Legal Workspace
Explore legal sources in Word OpenLegalCore Word Connector · public source-code beta

02Integration into an organisationOLC Engine 0.1.0 · private · active development

The legal method within your workflow

We can integrate Engine and the relevant legal foundation into your application or a specific step in your workflow. Together, we define the scope of implementation, connections to existing systems and how material is handled. Under a partnership agreement, we can adapt the working environment to your organisation and brand.

Explore ways to use OpenLegalCore

System

Four parts.One connected way of working with law.

OLC Engine connects three groups of components: legal sources, document processing, and workspaces and connectors. The groups differ in purpose, development stage and access terms.

01

Legal core

Engine

0.1.0 · private · active development

The legal core connects case material, legal sources, analysis and drafting in a traceable workflow.

02

Legal foundation

Legal sources

Slovenian legislation and case law · private datasets

Two prepared legislation and case-law datasets. Data access is private and arranged by agreement.

03

Document preparation

Document processing

OCR and two pipelines: public source code · fact extraction: private

OCR prepares documents for review, two pipelines acquire and maintain legal sources, and fact extraction remains a private component.

04

Work and connectors

Workspaces and connectors

Workspace and Studio: in development · Word Connector: public source-code beta

Legal Workspace connects matters and documents, Studio supports document and speech work, while Word Connector brings legal questions into Word.

Why open / why this core exists

Legal AI must make its work verifiable.

Open code is only the beginning.Legal infrastructure must preserve an inspectable record from source to use: what entered the system, how it was transformed, which limits apply, what human review covered and who remains responsible.

The standard /

An answer is not enough.

For OpenLegalCore, a legally material result is incomplete unless its sources and transformations can be traced, its method reviewed, its stated limits challenged, its human-review status recorded and responsibility for its use kept explicit. That inspectability is an engineering requirement—not a claim that open software makes the result correct.

The accountability record / five elements

  1. Source

  2. Transformation

  3. System limits

  4. Human review

  5. Responsibility

Across the record / four obligations

Four obligations turn openness into an operating discipline.

  1. Evidence

    Inspect before trust

    Code, methods, source handling and the scope of each claim must remain examinable within their stated access boundaries. Openness does not prove correctness; it makes unsupported claims easier to detect.

  2. Control

    Control the operating boundary

    Institutions must be able to choose their infrastructure, models and deployment boundary instead of making legal documents and system behaviour dependent on one opaque service.

  3. Infrastructure

    Build reusable infrastructure

    We build infrastructure that can be inspected, tested and developed. Public document-processing and ingestion tools, private legal datasets and Engine have distinct roles and terms of use. OCR and fact extraction also support work within Engine.

  4. Responsibility

    Keep responsibility visible

    Legally material output must remain traceable, reviewable and contestable. A system may support a decision; the person or institution making it remains responsible.

Required outcome /

Every legally material claim must remain connected to its sources, transformations, stated limits, review status and responsible person or institution.

Inspect component records

SYSTEM CORE / PRIVATE · 0.1.0

OLC Engine

The cognitive architecture of legal work.

Version 0.1.0 focuses on Slovenian law. Engine connects maintained legal sources, document processing, persistent Case Memory, analysis, strategy, legal drafting and review. The core remains private; commercial licensing and terms of use are agreed individually with each partner. Human review remains essential.

Open the engine record

REFERENCE ARCHITECTURE / SIX FUNCTIONAL LAYERS

The diagram shows the connected functions of Engine 0.1.0.
Maintained legal sources → controlled legal work
  1. 01

    Legal sources

    PISRS, Slovenian case law and controlled matter documents.

  2. 02

    Matter context

    Persistent Case Memory across Conversations and Matters.

  3. 03

    Legal functions

    Rules, retrieval and specialist legal functions.

  4. 04

    Analysis + strategy

    Legal analysis and comparison of available paths.

  5. 05

    Controlled drafting

    Legal drafting connected to the analytical foundation and review.

  6. 06

    Client or process

    OLC Legal Workspace or an embedded Engine service.

Case foundation

Documents and memory

  • Document processing, OCR and structured extraction
  • Persistent Case Memory
  • Source-linked assertions and evidence

Legal reasoning

Analysis and strategy

  • Legal analysis
  • Procedural strategy
  • Specialist legal functions

Legal work product

Drafting and review

  • Task-appropriate legal drafting
  • Review of the result and its foundation
  • Human responsibility remains explicit

The legal method where your work happens.

We can integrate OpenLegalCore into your legal department’s work or a selected organisational workflow. We start with a specific task: drafting or reviewing a contract, handling a claim, assessing a legal question or drafting a decision. The responsible team must review the material and results.

Legal work in different organisations

Companies with legal departments

Contracts and legal opinions

Reviewing contractual questions and drafting contracts and legal opinions. Material may include contract versions, supporting documentation and the company’s internal rules. The scope of analysis is tailored to the specific task.

Banks and financial institutions

Contractual relationships and complaints

Legal assessment of contractual relationships, complaints or a specific regulatory question. We connect the matter documentation and the institution’s internal rules with the relevant legal sources.

Hospitals and healthcare institutions

Procurement, suppliers and employment relationships

Legal questions involving public procurement, supplier contracts and employment relationships. For the selected matter, we define the question, documentation and material needed by the legal department or management.

Insurers and law firms

Compensation claims

Handling a compensation claim brings together documentation, relevant rules and legal assessment. For the selected workflow step, we define the scope of analysis, drafting and review of results.

Public administration and public institutions

Draft decisions

Drafting decisions and other legal material within a defined procedure. The sources and material underlying the draft remain identifiable. Review of the draft and responsibility for the final decision remain with people.

Universities and research institutions

Research collaboration and licensing

Reviewing agreements on research collaboration, licensing and rights to research results. We consider contract and project documentation and the relevant legal sources. The institution’s responsible department reviews the drafts.

These examples illustrate possible uses; they are not client references or evidence of completed production deployments. We agree the scope of implementation according to the task and available components.

Your documents.Your workflow.

Material, sources and access

We connect your documents, previous matters, internal rules and databases to OLC Engine and the relevant legal sources.

We prepare material for processing with AI, organise data connections and search, and define who may access each part of the material.

Workflow, interface and review

We adapt implementation to your systems and workflows. OLC can run within an existing system’s backend; reviewing results remains part of your work.

If the task needs an interface, we can choose , , a Word add-in tailored by agreement, or another existing or purpose-built interface.

· Public source-code beta

Components can be combined or integrated individually, including without a new interface. We agree the scope and terms of use.

How we connect material, legal sources and your systems

Conceptual illustration: documents and legal sources connect to OLC Engine, possible interfaces, organisational systems and human review.

← Swipe horizontally to explore the illustration →

A dashed connection marks planned source ingestion. For each task, we assess the sources’ legal status, terms of use and relevance.

What we define before implementation

Integration and access

We can integrate selected OpenLegalCore components, tools and legal data collections into an organisation’s business processes, individually or in combination, and connect them to existing systems. Integration and access arrangements are tailored to the organisation’s needs, taking into account the availability and licensing terms of the selected solutions.

We agree the scope of legal sources and search, data flows, how to connect to the model provider, human review of results and responsibility for operation.

Security depends on the whole operating environment:

  • access controls
  • updates
  • protection of login credentials and other secrets
  • backups
  • network isolation
  • how the system connects to the model provider

Technical connections

Technical integration can connect datasets, document stores, relational and vector databases, and systems for preparing and searching data.

OpenLegalCore

Legal AI / traceability contract

A legal answer needs a chain of custody.

A legally material result is incomplete unless the system can show its authoritative source, applicable version, retrieved passage, model and provider route, stated limits and human-review status. Together, these elements form its chain of custody.

Reference model
Source layer
GOVERNMENT WEBSITE / GOVERNMENT OFFICE FOR LEGISLATIONPISRS — Legal Information System of the Republic of SloveniaA government website providing access to legislation, case law and other legal information from Slovenia and the EU. It is administered by the Government Office for Legislation. Unofficial consolidated texts make legislation easier to consult, but are not official publications. PISRS also provides access to authentic electronic editions of the Official Gazette of the Republic of Slovenia from 1 March 2026.Open PISRS · OFFICIAL SOURCE / SUPREME COURT OF THE REPUBLIC OF SLOVENIASlovenian case lawSlovenia's authoritative national public case-law collection, published through the judiciary's official service administered by the Supreme Court of the Republic of Slovenia. It is the complete public corpus available through that service; statutory publication rules do not require every decision issued by every Slovenian court to appear.Open the official collection · controlled documents
Time boundary
Applicable text at the relevant date
Retrieval
Passages with stable source identifiers
Engine
Jurisdiction-specific legal orchestration
Interface
Legal Workspace or embedded service
Responsibility
Human review remains explicit

Required behind every legally material output

  1. 01Jurisdiction + matter boundarynamed
  2. 02Source + applicable versionretained
  3. 03Retrieved passage + identifierinspectable
  4. 04Model + provider routedisclosed
  5. 05Limits + uncertaintyvisible
  6. 06Human-review statusexplicit

How we are developing OpenLegalCore

Engine already brings together legal method, case material and Slovenian legal sources. We are developing Legal Workspace, broadening the legal foundation and preparing to use the system in other jurisdictions.

View the full roadmap
  1. Today

    Engine 0.1.0

    From clarifying the task to analysis, procedural strategy and legal drafting. Engine uses case material and persistent case memory, and includes OCR, pagination and document import.

    It connects a universal method of legal work with the sources and rules of a specific jurisdiction and the material of an individual case. The foundations, results and open questions of that work can be reviewed.

    OLC Engine is a privately developed legal core. Licensing and terms of use are agreed individually with each partner.

  2. Today

    The Slovenian jurisdiction

    Slovenian legislation and case law already form Engine's legal foundation. We acquire, curate and regularly maintain the sources in datasets, preserving links to the original records.

    For work with Slovenian law, Engine uses legislation and case law. We build this legal foundation by regularly acquiring, curating and maintaining sources. We preserve links to the original records and information about legislation's versions over time.

    We prepare the acquired material for use in datasets and search. Curation includes organising texts and document metadata, preserving source references, and checking the technical consistency of the data and whether refreshes have succeeded.

    The software tools are publicly available under BUSL-1.1. Partnership terms apply to access to the prepared and maintained datasets.

  3. In development

    Legal Workspace

    Conversations, matters, files and notes in a working environment connected to Engine. We are developing it for work in a personal space and collaboration among members of an organisation.

    Status
    in development.

    We are developing Legal Workspace as an environment for conversations, matters, files and notes, and for working with Engine. It is designed for work in a personal space and collaboration among members of an organisation.

    Our focus is on preparing an environment for day-to-day legal work, where an individual answer can form part of the broader work on a matter.

  4. EU law and additional legal sources

    We plan to include EU law and the case law of the Court of Justice of the EU and the European Court of Human Rights. The expansion also includes searching the publicly available decisions and opinions of the Information Commissioner.

    Planned release
    November 2026.

    We plan to broaden the legal foundation with European Union law and the case law of the Court of Justice of the European Union and the European Court of Human Rights. This expansion will also include searching the publicly available decisions and opinions of the Slovenian Information Commissioner.

    We will connect the individual groups of sources with Engine's legal method, taking account of their origin, the appropriate version over time and their role in legal judgment.

    These periods are development targets. When the releases are ready, we will also publish information about their availability and terms of use.

  5. In testing

    Other jurisdictions

    Engine's cognitive method remains the same. For other jurisdictions, we are preparing local legal sources, establishing the rules governing their use and checking the results.

    Status
    in testing.
    Planned release
    December 2026.

    Engine's cognitive method remains the same, while the legal foundation is adapted to each jurisdiction. Local sources, their versions over time and the rules governing their use need to be established for each one.

    These periods are development targets. When the releases are ready, we will also publish information about their availability and terms of use.

Governance / Operating truth

Open code does not meanhidden responsibility.

OpenLegalCore begins with founder-led governance. Public discussion is welcome; product direction, access, releases and the project identity retain clear accountability.

01

Public foundation

Useful on its own.

Working software, documentation, tests, synthetic examples, evaluation tools and review methods are released only after the relevant readiness gate.

No deliberately crippled teaser repositories
02

Private incubation

Protected while incomplete.

Unfinished components, sensitive integrations and production evidence remain private until rights, limits and release conditions are resolved.

No promised opening date
01 / MODEL

Founder-led

Maintainer rights are earned through sustained judgement and trusted work.

02 / LICENCE

Component decisions

Code, data, documentation and brand rights are handled separately. Each component’s licence and access terms are stated in its own record.

03 / PRIVACY

Synthetic public data

Private legal documents, evidence packages and credentials stay outside public repositories.

04 / SECURITY

Private reporting path

Suspected vulnerabilities now have a monitored project-wide private channel. Public issue trackers must not be used for security reports.

Report privately

Funding does not buy governance, maintainer status or automatic access to private source code.

Articles / Long-form reading

Methods, architecture and arguments.

Long-form work on verifiable legal AI, public methods, system design and the limits of automation.

View all articles

OPENLEGALCORE / CORE ARGUMENT

Legal AI infrastructure

Who Controls the Legal AI Stack?

Public code can enable inspection without proving control of the whole system. Follow versions, provenance, data sovereignty, exit and regulatory roles in legal AI.

Author
Rajko Majcen
Reading time
37 min
Research verified
30 October 2026
Read core argument

LATEST PUBLICATIONS

  1. 01

    Law-firm data security

    A Server in the Office Is Not a Security Strategy

  2. 02

From public record to shared work

Build the legal infrastructurethat should exist.

Tell us what you want to achieve. We can discuss using or integrating the system, contributing expertise or development work, and partnerships or funding.

01 / Expertise and development

Contribute the expertise the work needs.

Describe the problem and how you could contribute. Code, tests, documentation, review of the legal method and knowledge of sources are good starting points. You can also contribute directly in the public repositories.

Public repositories

02 / Use and partnerships

Let’s talk about your goal.

Tell us about your idea for using the system, a pilot or integration, adaptation for your organisation, or a partnership or funding opportunity. A goal and an outline of the scope are enough to begin.

Access boundaryPrivate access is scoped to agreed work. It does not create ownership, maintainer status or decision rights.

A concrete problem is enough to begin.

Project accountability

Responsibility has a name.

Open participation broadens the work without blurring who is accountable for project direction, releases and public–private boundaries.

Rajko Majcen, founder and maintainer of OpenLegalCore.

FOUNDER · MAINTAINER

Rajko Majcen

Rajko Majcen is a technology founder, technical project leader and product builder with more than three decades of work across security engineering, electronic banking, telecommunications, enterprise systems, open-source infrastructure and applied AI. He founded OpenLegalCore to bring that systems experience to transparent, verifiable legal technology.