Cyborg has announced the availability of its Enterprise RAG Blueprint, a solution designed to bring full encryption-in-use to enterprise-grade retrieval-augmented generation (RAG) workflows.
The blueprint is now available on build.nvidia.com and GitHub. It enables organizations to deploy secure RAG workflows and vector embeddings utilizing the CyborgDB encrypted vector store. This is achieved while maintaining performance, powered by NVIDIA Nemotron open models, NVIDIA NeMo Retriever microservices, and NVIDIA accelerated computing.
Nicolas Dupont, Founder and CEO of Cyborg, stated, “Today’s organizations want to unlock value from AI by centralizing their knowledge into a single vector database to make models more capable and context-aware. That consolidation is fundamental, but it also creates a smaller attack surface with a much larger potential breach radius. Vector databases can therefore become an organization’s biggest liability or its greatest strength. Encryption-in-use addresses this paradox by enabling enterprises to embrace AI confidently without turning innovation into exposure.”
Organizations such as OWASP have indicated that vectors and embeddings are an emerging area of vulnerability. While traditional vector databases encrypt data at rest and in transit, they typically process queries in plaintext. The Cyborg Enterprise RAG Blueprint employs a different method by eliminating the potential exposure of sensitive information through full encryption-in-use. This ensures that plaintext data does not exist in memory, logs, caches, or during search operations.
The Cyborg Enterprise RAG Blueprint operates through several key steps. For Embedding Generation & Cryptographic Indexing, user data is parsed and converted into embeddings using an NVIDIA NeMo Retriever embedding model. These embeddings are then cryptographically indexed via CyborgDB, producing encrypted tokens that are stored in standard backing stores with vector search capabilities, such as Redis or PostgreSQL. During Encrypted Retrieval, prompts are embedded and sent to CyborgDB for cryptographic retrieval, with a NeMo Retriever reranking model optimizing results for accuracy and quality. The system also incorporates forward-secure indexing to prevent reconstruction attacks on historical data. Key Management involves a launchable notebook that generates an encryption key once and stores it in base64 format on disk, with enterprises retaining full control and ownership of these encryption keys.
By integrating NVIDIA NIM microservices and NVIDIA cuVS GPU-accelerated search with CyborgDB’s encryption-in-use, the Cyborg Enterprise RAG Blueprint aims to deliver data protection without compromising enterprise-grade performance. The architecture supports multimodal capabilities, including PDF parsing, advanced table and chart extraction, hybrid search, and reranking with NVIDIA NeMo Retriever, achieving sub-10ms encrypted query performance.
Deployment guides for the Cyborg Enterprise RAG Blueprint are available on build.nvidia.com, enabling users to deploy the complete solution with CyborgDB’s encrypted vector indexing and retrieval. System requirements include a Docker deployment with a minimum of 2x NVIDIA H100 or 3x NVIDIA A100 GPUs, or a Kubernetes deployment requiring 8x H100-80GB or 9x A100-80GB. Alternatively, users can utilize NVIDIA NGC-hosted NIM with 1 NVIDIA GPU for CyborgDB acceleration. The operating system requirement is Ubuntu 22.04.
Included in the blueprint are NVIDIA AI software components such as NeMo Retriever and Llama Nemotron 3.3, CyborgDB with NVIDIA cuVS GPU acceleration, NeMo Retriever multimodal PDF parsing, NeMo Guardrails, OpenAI-compatible APIs, a sample UI, and production deployment configurations for both Docker and Kubernetes. Additional information is available on cyborg.co.
Cyborg focuses on digital privacy within enterprise AI through its flagship product, CyborgDB. CyborgDB functions as a vector database proxy that provides full encryption-in-use, maintaining the encryption of vectors, metadata, and keys at all stages. Designed for compatibility with existing databases, CyborgDB enables organizations to develop and scale AI systems while aiming to ensure security and compliance.
