Project-Based Visual Ontology Authoring – Provides an intuitive SHACL-first environment for defining classes, properties, relationships, and validation constraints without relying on complex ontology syntax.
OWL-to-SHACL Transformation & Standards Interoperability – Imports existing OWL/RDF models and transforms them into practical SHACL structures, helping organizations reuse established enterprise and industry standards.
Modular Ontology Management & Reference Reuse – Enables teams to extend shared, read-only reference ontologies while keeping local models modular, governed, and easier to maintain.
Human-in-the-Loop Review Sandbox – Stages AI-generated changes for visual review and approval before they are committed, keeping experts in control of the active model.
Baseline Versioning & Enterprise Catalog Preview – Adds immutable model snapshots, restore capabilities, and the foundations for centralized discovery and reuse of enterprise semantic models.
Unified AI Taxonomy Enrichment & Taxonomy Builder Overhaul – Simplifies AI-assisted curation with focused in-context concept enrichment and a more flexible Taxonomy Builder supporting multi-selection, direct workflow entry, and flat or hierarchical concept generation.
Bulk Operations via Multi Drag & Drop – Lets curators move, re-parent, add, or use multiple concepts in a single action, significantly reducing repetitive work during taxonomy restructuring and collection management.
Unified Single Sign-On (SSO) – Extends SSO to MirrorApp, GraphViews, and Semantic Workbench, providing seamless cross-tool access and more centralized identity, MFA, and access governance.
Tagger API Performance & Resilience Optimization – Streams tagging results, handles large asynchronous payloads more efficiently, and allows batch processing to continue when individual documents fail, reducing memory pressure and improving stability at scale.
Explicit Lucene Index Refresh via API – Enables teams to rebuild and warm extraction indexes after taxonomy or data updates, ensuring up-to-date tagging results and more predictable response times.
AI Evaluation & Benchmarking – Introduces a self-contained evaluation framework using RAGAS metrics, multi-provider LLM judging, actionable HTML reports, and automated baselines for repeatable quality and regression testing.
Extended Concept Details – Adds full support for taxonomy attributes, custom properties and relations, with clickable navigation between broader, narrower, and related concepts.
Stronger Authentication & Security – Replaces Basic Auth with OAuth2 for PoolParty integrations, moves n8n API keys into managed credentials, and strengthens the security of workflow infrastructure.
Improved GraphRAG User Experience – Adds collapsible conversation subthreads, clearer WCAG AA-compliant status messaging, and refreshed interaction text for a more usable interface.
Incremental Taxonomy Sync & Term Integrity – Replaces full taxonomy resyncs with delta-based updates and safely handles hierarchy changes, term consolidation, and polyhierarchy while preserving stable identifiers, existing tags, and search behavior.
Governance-Aligned Autotagging – Ensures only approved, published concepts are synchronized and available for tagging, with lifecycle changes automatically reflected across connected content repositories.
Multilingual Autotagging & Search – Improves concept detection, tagging, and search consistency for non-English content, strengthening support for global and multilingual Microsoft 365 deployments.
Simplified Deployment & Operations – Reduces configuration and upgrade overhead, making deployments easier to maintain, while incremental synchronization lowers the operational cost of keeping large taxonomies aligned.
Ready-Made GraphDB Components – Provides pre-built components for GraphDB operations, data modeling, transformation, and semantic analytics, reducing the need for custom API integration code.
Content Ingestion for GraphRAG – Introduces reusable document parsing, chunking, and loading workflows for formats such as PDF, DOCX, and TXT, replacing bespoke ingestion pipelines with configurable steps.
Professional Services Task Runner – Gives delivery teams a supported environment for running custom Python inside the platform instead of maintaining scripts externally.
Deployment Foundations – Establishes the Docker image, system architecture, development environment, and operational documentation needed for repeatable enterprise deployments.
Governed Full & Incremental Taxonomy Sync – Keeps AEM Tag Manager aligned with Graph Modeling through full baseline synchronization and targeted incremental updates, using concept URIs to preserve existing content references and avoid duplicate tags.
Semantic & Governance Preservation – Carries SKOS metadata, hierarchy, multilingual labels, editorial status, and custom properties into AEM while enforcing Graph Modeling as the authoritative source and safely handling renames, moves, archives, and merges.
Enterprise Scale, Security & Auditability – Supports AEM as a Cloud Service and AEM 6.5, introduces OAuth 2.0 or API-key authentication, provides detailed audit logging, and is designed for taxonomies of 50,000+ concepts with headroom to 100,000.
Auto-Tagging Foundation – Establishes annotation-location registration and polling capabilities so external enrichment pipelines can identify changed AEM content and support future automated tagging workflows.
Repository-Specific Maintenance Permission – Enables administrators to grant management capabilities for individual repositories without giving users global repository management privileges, improving security and access control.
Multiple Authorization Sources – Allows LDAP, OAuth, and local authorization to work together, making it easier to grant temporary repository access without changing centrally managed identities or groups.
Unified Prometheus Metrics Endpoint – Consolidates instance, repository, cluster, query, and license metrics into a single endpoint, simplifying production monitoring and reducing configuration overhead.
Native Index Fragmentation Insights – Exposes detailed repository index statistics, including fragmentation levels, helping administrators determine when compaction is needed to maintain efficient operations.
Encryption at rest (Enterprise Edition): Secure sensitive C3/C4 data with repository-level AES-256-GCM encryption, configurable at repository creation to seamlessly meet enterprise compliance standards.
Interactive Map Visualizations: Visualize geospatial SPARQL results (points, lines, and polygons) directly on an interactive, embeddable map component inside the GraphDB Workbench.
Custom MCP Prompts: Define Model Context Protocol (MCP) prompt templates via REST API, allowing AI agents to safely translate natural language into optimized SPARQL queries.
Unified Platform Security & SSO: This release transitions the entire platform to a decentralized Keycloak architecture to enable seamless Single Sign-On (SSO) across the Graphwise suite. Key upgrades include application-specific role isolation to prevent token bloat, secure machine-to-machine API integrations, and an automated zero-downtime migration utility.
Taxonomy Builder: Multilingual Localization: This feature integrates advanced LLM technology to automatically generate concept labels and definitions across all configured project languages directly within the application workflow. It drastically reduces manual translation costs and timelines for global enterprises while maintaining human-in-the-loop quality controls.
Native Multilingual Engine: This Apache Lucene-based upgrade enables high-performance concept tagging across more than 40 languages, introducing native support for complex scripts like Japanese and Russian out of the box. By utilizing memory-efficient technology, it eliminates the need for external Elasticsearch lookups and fully automates the indexing process whenever a taxonomy changes.
Unified Tagger API: This feature consolidates multiple legacy extraction, annotation, and categorization endpoints into a single, high-performance API (/api/tag). It streamlines multi-document processing and ZIP file handling into a single request, significantly reducing integration complexity for developers.
Semantic Explainability: Integrates richer SKOS attributes into backend responses and the UI to provide deeper semantic context. This significantly enhances transparency and answer explainability through better concept grounding.
Guardrails & Trust Layer: Upgrades output guardrails to provide descriptive explanations for off-topic questions while suppressing irrelevant sources. Auto-generated follow-up questions are also refined to better match the active conversation context.
Graph & Workflow Reliability: Uses a custom MCP client to intelligently route queries to vector search or GraphDB instead of running them in parallel. This drastically reduces latency and token usage for simple queries while stabilizing enterprise workflows.
On-Demand Retagging: Users can now trigger tagging manually for an entire SharePoint document library or for selected individual documents. This provides more control when metadata needs to be refreshed after taxonomy updates, bulk migrations, major document revisions, or other changes outside the regular delta tagging cycle.
Faster Sync Performance: Connectors now uses the new extractor, improving synchronization speed and efficiency, especially for larger SharePoint libraries and environments with frequent content updates.
Improved Taxonomy Navigation: The manual sync interface now displays taxonomy hierarchies with clearer indentation, making parent-child relationships easier to understand at a glance. This improves usability for users reviewing or working directly with structured vocabularies during manual sync activities.
