Verbis Graph Engine – GraphRAG Knowledge Retrieval
بواسطة Baida ALena
GraphRAG on Azure Kubernetes for vector search, knowledge graph traversal and AI output validation,
Verbis Graph Engine is a GraphRAG knowledge retrieval application for private document collections. It combines semantic vector search with knowledge graph traversal to support reliable AI retrieval, source-grounded answers, AI output validation, and multi-hop reasoning across documents, entities, and relationships.
Delivered as a Kubernetes application for Azure Kubernetes Service (AKS), Verbis gives developers and technical teams a practical path from a free GraphRAG evaluation to ongoing PAYG usage through the same Microsoft Marketplace offer.
Key capabilities
- GraphRAG hybrid retrieval: combine vector similarity with knowledge graph relationships for connected, context-aware results beyond traditional vector-only RAG.
- Private document retrieval: index and query PDF, DOCX, CSV, and TXT content across multiple sources.
- AI output validation: inspect cited answers, connected evidence, and source material instead of relying only on generated text.
- Knowledge graph exploration: discover relationships between entities and information distributed across documents.
- Developer access: use the web interface or FastAPI endpoints, with OpenAPI documentation available at
/docs. - Operational readiness: use the health endpoint at
/_stcore/healthfor orchestration and monitoring.
Free Evaluation
Evaluate GraphRAG, vector search, and knowledge graph traversal at no software license cost. The Free Evaluation plan is designed for demonstrations, learning, integration tests, and small document collections. It is not intended for large production workloads. Azure infrastructure costs still apply.
PAYG per pod
The commercial PAYG plan is billed through Microsoft Marketplace according to active pod usage and the price shown for the plan. It is designed for ongoing pilots and operational workloads without an upfront capacity commitment.
Deployment and data control
- Installs on a customer-managed AKS cluster through the Azure Marketplace cluster-extension experience.
- Runs inside the customer Azure environment.
- Customers control AKS sizing, network policy, storage configuration, and Azure infrastructure costs.
- Persistence depends on the storage configuration selected for deployment; container-local data should not be treated as durable across pod replacement.
- External model services are optional and depend on the selected configuration and plan.
Typical use cases
- Enterprise knowledge assistants and private AI search
- Technical documentation and internal support
- Research and multi-document analysis
- AI governance, auditability, and compliance workflows
- Due diligence and regulated knowledge bases
Verbis does not use customer content to train shared models. See the privacy policy and product documentation linked in this listing for operational and data-handling details.