OPAQUE Unveils Confidential Agents with Turnkey RAG Workflows, Reinventing Trust and Security in AI

OPAQUE Unveils Confidential Agents with Turnkey RAG Workflows, Reinventing Trust and Security in AI

‍

San Francisco, CA – June 17, 2025 – OPAQUE Systems, the leader in Confidential AI, today announced at the 3rd annual Confidential Computing Summit the launch of its Confidential Agents for Retrieval-Augmented Generation (RAG with turnkey workflows built on NVIDIA NeMo Guardrails and LangChain’s agent framework, LangGraph). This groundbreaking solution transforms AI adoption by enabling enterprises to rapidly and securely deploy sophisticated, confidential AI agents.

Unlike traditional agent frameworks, OPAQUE Confidential Agents provide infrastructure-level confidential guarantees, ensuring sensitive data remains encrypted at every step within hardware-backed Trusted Execution Environments (TEEs). Enterprises can confidently enhance their agentic workflows with built-in runtime policy enforcement and verifiable auditability across every agent action, seamlessly integrating these confidentiality guarantees into their existing AI frameworks.

“At Bloomfilter, we believe the future of software delivery is open, observable, and trustworthy." said Erik Severinghaus, Co-Founder and Co-CEO of Bloomfilter. "OPAQUE enables us to build AI-native systems where data confidentiality isn’t an afterthought—it’s foundational. Their Confidential Agents give us the infrastructure to scale with speed and clarity, without sacrificing trust. It’s the rare case where security and velocity reinforce each other.”

OPAQUE's platform is uniquely designed to:

• Secure Every Interaction: Actions are executed within attested, hardware-backed TEEs, maintaining data encryption during retrieval and inference.
• Ensure Real-Time Compliance: Policies and guardrails, defined through NVIDIA NeMo Guardrails' Colang policy language, are continuously enforced, providing real-time cryptographic governance.
• Deliver Complete Auditability: Detailed audit logs and cryptographic proofs provide verifiable transparency into every agent decision, which is critical for all industries that use sensitive data, including financial services, fintech, and enterprise B2B SaaS.
• Enable Multi-Agent Collaboration: Supports sophisticated workflows where agents escalate, delegate, and adapt autonomously within secure environments, facilitating complex, enterprise-grade deployments.

“OPAQUE is accelerating innovation velocity by merging powerful AI agent technologies with robust, verifiable guarantees for sensitive data,” said Harrison Chase, Co-Founder and CEO of LangChain. “As enterprises increasingly orchestrate agentic workflows into complex systems across customers, partners, and competitors, OPAQUE raises the bar—enabling agility, security, and trust at scale.”

OPAQUE Confidential Agents support all LangGraph agents, allowing customers the flexibility to leverage existing investments and easily integrate into their current AI ecosystems without additional complexity or replatforming.

For more information, please visit opaque.co/confidential-agents-for-rag.

About OPAQUE Systems
OPAQUE is the Confidential AI company. Rooted in groundbreaking research from UC Berkeley's RISELab, OPAQUE Systems delivers a Confidential AI platform that transforms how enterprises handle, analyze, and share sensitive data. Its Confidential Agents platform enables enterprises to run AI on sensitive data—without risk, delay, or compliance tradeoffs. Customers include ServiceNow, the European Union, and other global leaders in AI, fintech, and enterprise infrastructure. To learn more, visit opaque.co or request a demo to see Confidential AI in action.

Contact: Inkhouse for OPAQUE[email protected]