Stop Leaking, Start Building | Klave Newsletter October 2025
NEWS
Klave AI: Smarter, Faster, and Fully Secure
Klave AI empowers organisations to harness advanced AI without compromising data privacy. Whether using open-source or proprietary models, your data remains encrypted and fully under your control, even during processing.
With Klave AI, you can:
Confidently interact with AI while keeping information private
Securely access private documents using Retrieval-Augmented Generation
Receive precise, context-aware insights without sacrificing confidentiality
Built for organisations that demand both intelligence and security, Klave AI ensures your data, models, and insights remain fully protected.
Watch the demo to see Klave AI in action and discover how it delivers secure, contextually aware AI.
Agentic AI and Secure Foundations
As organisations continue to embrace automation, the focus is shifting from what AI can do to how it can be implemented responsibly. The challenge lies in achieving efficiency and intelligence without compromising data protection or transparency.
Klave AI addresses this challenge head-on. It is designed to uphold data sovereignty, ensuring that sensitive information remains private and verifiable at all times. Through its confidential computing framework and secure-enclave attestation, Klave AI delivers trust and accountability at the core of every AI interaction.
Our platform is also fully MCP compatible, allowing seamless integration with modern AI ecosystems while maintaining the highest security standards.
To see these principles in action, we have released a short demonstration of Klave AI in a practical use case, showcasing how secure infrastructure and contextual intelligence work together: Watch the preview on LinkedIn
Our next milestone focuses on enabling MCP servers to run directly within secure enclaves, further strengthening the privacy and reliability of agentic AI deployments.
TECH
Klave v3.6.0 Release
The latest update, v3.6.0, brings improvements in security, speed, and usability across the platform.
Enhanced Cryptography: New methods in the Crypto SDK, including derive_bits and derive_key with ECDH and HKDF, are now available in Rust and AssemblyScript. This allows for more secure and adaptable key management workflows.
Performance Upgrades: Optimised HTTP outcalls improve network efficiency, enabling faster AI operations and smoother system interactions.
Improved User and Developer Experience: Interface responsiveness has been enhanced, workspaces are more intuitive, and components are more reactive for seamless navigation.
Preparing for the Future:
Direct HTTP connections for simplified app integration while maintaining Remote-Attested TLS security
Foundations for Klave Single Node to deploy AI inference, agents, RAG, and MCP servers more easily
Support for full-chain attestation (SGX, TDX, NVIDIA), enabling fully verifiable AI workflows within secure enclaves
Explore the full capabilities of v3.6.0 at klave.com and review our updated documentation.
NVIDIA Inception Program
We are proud to join the NVIDIA Inception Programme, gaining access to the latest GPUs, tools, and industry networks.
Klave AI protects sensitive data and models through Confidential Computing, keeping data encrypted and isolated even during processing. From secure semantic search to AI agents powered by MCP, our platform enables safe automation across private data sources in regulated environments, ensuring data is never exposed or used for training without consent.
For a tailored demonstration of Klave AI, please contact us directly.
FEATURED ARTICLES
Discover the New Reads
Why You Should Switch to Klave AI
Written by Étienne Bossé, Head of Platform at Secretarium.
Klave AI provides a cryptographically secure platform that protects every stage of modern AI workflows, keeping data, credentials and model IP safe even during use. As centralised LLMs pose risks of data leakage and loss of control, Klave enables a hybrid architecture that maintains both high performance and strong privacy.
The CJEU Ruling: The Rise of Confidential Computing
Written by Abby Sawbridge, Legal & Business Ops at Secretarium.
This article explores the European Court of Justice’s landmark ruling on one of the most persistent questions under the GDPR: At what point does pseudonymised data stop being considered personal data? It reinforces what privacy leaders have long understood: confidential computing is essential for achieving secure and fully compliant data protection across Europe.
Written by Abby Sawbridge, Legal & Business Ops at Secretarium.
This article delves into the benefits of integrating Retrieval-Augmented Generation (RAG) with private data sources. The article discusses how this approach enhances AI performance while maintaining stringent data privacy and compliance with regulatory requirements.
Building Honest Systems with Rust, WebAssembly, and Confidential Computing
Written by Étienne Bossé (Head of Platform) and Jean-Jacques Lafay (Senior Software Engineer) at Secretarium.
This article presents insights from use of Rust and WebAssembly in creating secure, verifiable systems. The article highlights how these technologies contribute to building 'honest systems' that ensure data integrity and confidentiality in all environments.
If you’re interested in working with us, please email: contact@klave.com
If you like these monthly updates from Klave, feel free to share this link with your network. Our monthly newsletter can also be found via our LinkedIn page.
Feel free to browse our previous newsletter editions here.
 (1).jpeg)
