One architecture. Use only what you need.

A reusable AI-native OCR pipeline with deterministic controls, ordered artefacts and a public offline tool for page-level human verification.

Current boundaryThe first public OpenLegalCore component is published as v0.1.2 under Apache-2.0. Code, documentation, tests, offline review tooling and the canonical acceptance record are public; the source filing and generated acceptance artefacts remain private.

A functional first iteration that converts Slovenian legal text into source-linked candidate assertions and directives, designed as a reusable foundation for structured matter memory and later adaptation across legal workflows.

Current boundaryThe standalone extraction core exists. Open WebUI document-upload integration, OLC Legal Workspace integration and persistent structured matter memory are planned rather than current capabilities; extracted items remain review candidates, not canonical facts.

A working private legal-retrieval and source-bound answer foundation for the current Slovenian reference jurisdiction, being developed as OpenLegalCore's general-purpose legal cognition engine and system core.

Current boundaryVersion 0.0.7 supports bounded legal-source planning, maintained Slovenian legislation and case-law paths, admissible source selection, cited text or structured output, source validation and explicit completion states through a private OpenAI-compatible text interface. Approved organisations can use the current Engine in a scoped evaluation, integration or pilot after its purpose, deployment boundary, data access, support and terms are agreed through an individual partnership. This is not public source access or an anonymous service. Persistent Case Memory, production Document Intelligence, general complex-matter analysis, strategy, specialised drafting, deadline and jurisdiction tools, extended legal-source families, durable Analyst findings and controlled whole-file review are not part of this version.

A source-available ingestion and integrity-maintenance pipeline that discovers and structures PISRS legislation in authoritative PostgreSQL state and maintains a derived Qdrant semantic index.

Current boundaryVersion 0.1.0 is public under BUSL-1.1. Code, documentation, tests and deployment templates are public; PISRS credentials, source data, legislative text, PostgreSQL or Qdrant snapshots, managed infrastructure and retrieval or user interfaces are not.

A source-available pipeline that checks supported Slovenian case-law and related legal records, maintains source-linked PostgreSQL state and a derived Qdrant index for OLC Engine and compatible legal workflows.

Current boundaryVersion 0.1.7 is public under BUSL-1.1. The repository contains code, documentation, tests and deployment templates, but no source credentials, court decisions, source records, embedding access, managed infrastructure, retrieval interface or legal advice.

A carefully curated and regularly maintained Slovenian legislation data foundation for semantic search, legal portals, research workflows and OLC Engine, available as a private download to approved partners.

Current boundaryThe maintained PostgreSQL state and derived Qdrant index are offered only through a scoped data partnership. They are not published on GitHub or offered as an anonymous public download; PISRS and the official publication remain authoritative.

A carefully curated and nightly maintained Slovenian case-law data foundation for semantic search, legal portals, research workflows and OLC Engine, available as a private download to approved partners.

Current boundaryThe coordinated PostgreSQL and Qdrant package is refreshed after each successful nightly synchronisation and offered only through a scoped data partnership. It is not published on GitHub or offered as an anonymous public download, and the official source remains authoritative.

A planned first-party, matter-centred legal workspace across web and desktop clients, connecting uploaded or user-authorised local files with evidence, AI Notes, verifiable sources and controlled drafting.

Current boundaryThe product direction is recorded. The web client is planned first and desktop clients later; none is available. No public repository, release, filesystem scope, conferencing integration, offline mode or feature-complete client is claimed.

A public bilingual Microsoft Word connector that sends a question without document context or only explicitly selected text to OLC Engine—directly or through Open WebUI backed by OLC Engine—and returns source-linked answers without changing the document.

Current boundaryThe Apache-2.0 v0.1.0-beta.1 release is source-only beta software verified in Word for the web. It includes no hosted OpenLegalCore service, production backend, legal database, protected staging access, AppSource or Office Store listing, automatic provider fallback, credential sharing or document-editing workflow. Word for Windows and Word for Mac are not tested.