+ Schedule your live demo session with the Ultraviolet team.
+
+
+
+
+
+
+
+
+
+ Secured via Google Calendar
+
+
+
+
+
diff --git a/src/content/blog/amd-sev-snp-vs-intel-tdx.md b/src/content/blog/amd-sev-snp-vs-intel-tdx.md
index be1a4aa..ed63ecd 100644
--- a/src/content/blog/amd-sev-snp-vs-intel-tdx.md
+++ b/src/content/blog/amd-sev-snp-vs-intel-tdx.md
@@ -4,7 +4,7 @@ title: "AMD SEV-SNP vs Intel TDX: Understanding TEE Support in Cube AI"
description: "A technical comparison of AMD SEV-SNP and Intel TDX confidential computing technologies, and how Cube AI abstracts their complexities to provide a unified, secure platform for AI workloads."
excerpt: "A technical comparison of AMD SEV-SNP and Intel TDX confidential computing technologies, and how Cube AI abstracts their complexities to provide a unified, secure platform for AI workloads."
author:
- name: "sammy oina"
+ name: "Sammy Oina"
picture: "https://avatars.githubusercontent.com/u/44265300?v=4"
tags:
[security, confidential-computing, amd, intel, technical-deep-dive, "cube ai"]
diff --git a/src/content/blog/cocos-prism-release.md b/src/content/blog/cocos-prism-release.md
index 49e9ea1..55e706c 100644
--- a/src/content/blog/cocos-prism-release.md
+++ b/src/content/blog/cocos-prism-release.md
@@ -4,7 +4,7 @@ title: "Unleashing Confidential AI: Cocos v0.8.0 and Prism v0.6.0 Released"
excerpt: "Ultraviolet releases Cocos AI v0.8.0 and Prism v0.6.0, enhancing the confidential computing ecosystem. These updates bring improved attestation, stronger security, and a better enterprise experience."
description: "Learn about the new features in Cocos AI v0.8.0 and Prism v0.6.0, including enhanced security policies, performance improvements, and a new asset discovery system."
author:
- name: "sammy oina"
+ name: "Sammy Oina"
picture: "https://avatars.githubusercontent.com/u/44265300?v=4"
tags: [confidential-computing, ai, "cocos ai", "prism ai", privacy]
date: 2026-02-06
@@ -14,7 +14,7 @@ ogImage: /img/cocos-prism-release/cocos_prism_release_cover.png
featured: true
---
-The confidential computing landscape continues to evolve with two major releases that strengthen the ecosystem for secure, privacy-preserving AI: **Cocos AI v0.8.0** and **Prism v0.6.0**. Together, they represent a powerful advancement in enabling organizations to build trustworthy, collaborative AI systems without compromising data privacy.
+The confidential computing landscape continues to evolve with two major releases that strengthen the ecosystem for secure, privacy-preserving AI: **[Cocos AI](/products/cocos-ai) v0.8.0** and **[Prism AI](/products/prism-ai) v0.6.0**. Together, they represent a powerful advancement in enabling organizations to build trustworthy, collaborative AI systems without compromising data privacy.
diff --git a/src/content/blog/confidential-computing-meets-ai.md b/src/content/blog/confidential-computing-meets-ai.md
index 57eef6c..df1bd72 100644
--- a/src/content/blog/confidential-computing-meets-ai.md
+++ b/src/content/blog/confidential-computing-meets-ai.md
@@ -4,7 +4,7 @@ title: "Confidential Computing Meets AI: How Cube AI Protects Your LLM Prompts"
excerpt: "Cube AI leverages confidential computing and TEEs to protect your sensitive AI workloads. Ensure your prompts remain private with hardware-verified security."
description: "Discover how Cube AI leverages Trusted Execution Environments (TEEs) to protect your sensitive data and LLM prompts from cloud providers and insider threats."
author:
- name: "sammy oina"
+ name: "Sammy Oina"
picture: "https://avatars.githubusercontent.com/u/44265300?v=4"
tags: [security, confidential-computing, ai, privacy, "cube ai"]
image: /img/confidential-computing-meets-ai/confidential-computing-cover.png
diff --git a/src/content/blog/cube-ai-public-beta-launch.md b/src/content/blog/cube-ai-public-beta-launch.md
index 1e0ed4f..a798583 100644
--- a/src/content/blog/cube-ai-public-beta-launch.md
+++ b/src/content/blog/cube-ai-public-beta-launch.md
@@ -4,7 +4,7 @@ title: "Cube AI Public Beta: The Future of Confidential LLM Inference is Here"
excerpt: "We are incredibly excited to announce the public beta release of Cube AI. Unlock the power of open-source, hardware-secured generative AI. Build and deploy agentic workflows with absolute data privacy using Trusted Execution Environments."
description: "Discover Cube AI, the open-source framework by Ultraviolet for secure, agentic LLM inference. Learn how we use AMD SEV-SNP and Intel TDX to secure vLLM and Ollama deployments."
author:
- name: "sammy oina"
+ name: "Sammy Oina"
picture: "https://avatars.githubusercontent.com/u/44265300?v=4"
tags: [confidential-computing, ai, privacy, "cube ai", beta-launch, open-source]
date: 2026-03-04
@@ -19,7 +19,7 @@ The maturation of generative artificial intelligence has arrived at a structural
Organizations have been forced into a binary choice: sacrifice proprietary data and intellectual property to public cloud providers or forego the advantages of cutting-edge AI. Today, that choice is obsolete.
-We are incredibly excited to announce the **Public Beta of Cube AI**, an open-source framework developed by Ultraviolet that pioneers a "Confidential-by-Design" architecture. By leveraging hardware-based Trusted Execution Environments (TEEs), Cube AI secures the entire AI inference lifecycle.
+We are incredibly excited to announce the **Public Beta of [Cube AI](/products/cube-ai)**, an open-source framework developed by Ultraviolet that pioneers a "Confidential-by-Design" architecture. By leveraging hardware-based Trusted Execution Environments (TEEs), Cube AI secures the entire AI inference lifecycle.
@@ -33,7 +33,7 @@ Confidential computing addresses this through **Trusted Execution Environments (
Cube AI provides software enablement for leading TEE architectures: **[AMD SEV-SNP and Intel TDX](/blog/amd-sev-snp-vs-intel-tdx)**. By integrating with Buildroot and custom Linux kernels, we minimize the Trusted Computing Base (TCB), thereby significantly reducing the attack surface.
-To achieve this deep hardware integration, Cube AI relies on the open-source **[Cocos AI](https://cocos.ai)** framework—another flagship platform from Ultraviolet. Cocos AI provisions the secure **Hardware Abstraction Layer (HAL)** and manages the intricate **Remote Attestation** protocols. By leveraging Cocos, a client can cryptographically verify that the AI model is running inside a genuine hardware enclave with an unmodified software configuration. This establishes true "Zero Trust" at the hardware level.
+To achieve this deep hardware integration, Cube AI relies on the open-source **[Cocos AI](/products/cocos-ai)** framework—another flagship platform from Ultraviolet. Cocos AI provisions the secure **Hardware Abstraction Layer (HAL)** and manages the intricate **Remote Attestation** protocols. By leveraging Cocos, a client can cryptographically verify that the AI model is running inside a genuine hardware enclave with an unmodified software configuration. This establishes true "Zero Trust" at the hardware level.
## Architectural Innovation: The Cube AI Stack
diff --git a/src/content/blog/prism-public-beta-launch.md b/src/content/blog/prism-public-beta-launch.md
index 17cf973..bfda601 100644
--- a/src/content/blog/prism-public-beta-launch.md
+++ b/src/content/blog/prism-public-beta-launch.md
@@ -4,7 +4,7 @@ title: "Prism AI Public Beta: Empowering Secure AI Collaboration"
excerpt: "We are thrilled to announce the public beta release of Prism AI, our confidential computing platform for secure AI collaboration. Build and deploy privacy-preserving AI with hardware-verified security."
description: "Discover Prism AI, the web-based SaaS platform for secure, collaborative AI workloads. Learn how Prism AI leverages Trusted Execution Environments (TEEs) to protect sensitive data and algorithms."
author:
- name: "sammy oina"
+ name: "Sammy Oina"
picture: "https://avatars.githubusercontent.com/u/44265300?v=4"
tags: [confidential-computing, ai, privacy, "prism ai", beta-launch]
date: 2026-02-12
diff --git a/src/content/blog/vllm-vs-ollama-in-cube-ai.md b/src/content/blog/vllm-vs-ollama-in-cube-ai.md
index d449609..7b1de22 100644
--- a/src/content/blog/vllm-vs-ollama-in-cube-ai.md
+++ b/src/content/blog/vllm-vs-ollama-in-cube-ai.md
@@ -15,7 +15,7 @@ date: 2026-02-11
Selecting the right Large Language Model (LLM) backend is no longer just an infrastructure decision — it directly impacts latency, throughput, operational cost, scalability, and developer velocity. The choice of inference engine determines how quickly your models respond under load, how efficiently your GPU resources are utilized, and how much operational overhead your team must absorb on an ongoing basis. Getting this decision wrong can mean over-provisioning expensive hardware for a workload that doesn't need it, or under-serving users with an engine that can't keep up with production traffic.
-Cube AI is intentionally designed with **backend modularity** at its core, allowing teams to switch between inference engines without changing any application logic, rewriting API integrations, or modifying client SDKs. Whether you prioritize **GPU-accelerated performance** for high-concurrency production workloads or **lightweight local deployments** for development, edge computing, and confidential environments, Cube AI supports both paradigms through two production-ready backends:
+[Cube AI](/products/cube-ai) is intentionally designed with **backend modularity** at its core, allowing teams to switch between inference engines without changing any application logic, rewriting API integrations, or modifying client SDKs. Whether you prioritize **GPU-accelerated performance** for high-concurrency production workloads or **lightweight local deployments** for development, edge computing, and confidential environments, Cube AI supports both paradigms through two production-ready backends:
- **vLLM** — optimized for high-throughput, GPU-driven inference with continuous batching and explicit memory management
- **Ollama** — a flexible, developer-friendly runtime designed for local, hybrid, and resource-constrained environments
diff --git a/src/content/solutions/secure-collaboration.yaml b/src/content/solutions/secure-collaboration.yaml
index a10f608..aa75346 100644
--- a/src/content/solutions/secure-collaboration.yaml
+++ b/src/content/solutions/secure-collaboration.yaml
@@ -29,4 +29,4 @@ product:
ctaH2: "Collaborate without giving up your data."
ctaSub: "Talk to the team about Prism AI deployments, multi-party AI workloads, and free tier access."
ctaLabel: "View Prism AI"
-ctaHref: "/products/prism"
+ctaHref: "/products/prism-ai"
diff --git a/src/layouts/Layout.astro b/src/layouts/Layout.astro
index 2d31add..5887a13 100644
--- a/src/layouts/Layout.astro
+++ b/src/layouts/Layout.astro
@@ -32,7 +32,6 @@ const resolveMetaUrl = (value: string | undefined) => {
const canonicalUrl = resolveMetaUrl(canonical ?? Astro.url.pathname);
const ogImageUrl = resolveMetaUrl(ogImage);
-console.log("og image url: ", ogImageUrl);
---
@@ -58,11 +57,12 @@ console.log("og image url: ", ogImageUrl);
{ogImageUrl && }
{ogImageUrl && }
-
+
+
{ogImageUrl && }
@@ -92,6 +92,8 @@ console.log("og image url: ", ogImageUrl);
+
+