TraceCtrl LLM Guards (WebOrion Protector Plus)

CloudsineAI is a Singapore cybersecurity company based at Science Park, led by founder and CEO Matthias Chin. It started with web defacement monitoring and web application protection and now focuses on securing generative AI and agentic AI applications for government agencies, banks, healthcare providers, ISPs and universities.

TraceCtrl LLM Guards, formerly WebOrion Protector Plus, is a GPU-powered GenAI firewall that inspects prompts and model responses in API monitoring or forwarding mode. Its ShieldPrompt engine combines content moderation, system prompt protection, a threat vector database using semantic similarity and a CyberLLM model that detects malicious intent, plus keyword and regex filters and AI-generated red teaming prompts. It runs as an on-site hardware appliance, a cloud virtual appliance on AWS, Azure or Google Cloud, or as SaaS through API keys.

• Prompt injection defence: Blocks direct and indirect injection attacks and jailbreak attempts.
• Data leakage prevention: Detects PII, intellectual property and other sensitive data in inputs and outputs.
• ShieldPrompt engine: Combines content moderation, system prompt protection and a semantic threat vector database.
• CyberLLM: Uses a dedicated model to detect malicious intent in prompts.
• AI red teaming: Generates adversarial test prompts to probe LLM applications.
• Flexible deployment: Runs as a GPU appliance, cloud virtual appliance or SaaS via API keys.

A bank can place the firewall in front of its customer service chatbot to block jailbreaks and mask account data in responses. A hospital can protect a RAG assistant over clinical documents with on-site inspection. A university can let staff use internal GenAI tools while preventing confidential research from leaking.

Organizations can launch chatbots and RAG applications while blocking prompt injection, jailbreaks and system prompt leakage. Outbound inspection stops personal data and intellectual property from leaking through model responses. Three deployment options let regulated sectors keep inspection on site or scale elastically in the cloud. Controls are aligned with the OWASP Top 10 for LLMs, NIST AI RMF, MITRE ATLAS and Singapore CSA and GovTech guidance, which supports audit readiness.

Banks, public agencies and enterprises in Southeast Asia are rolling out GenAI chatbots and AI agents faster than their security controls, and regulators in Singapore and the region are publishing guidance on securing AI systems. Vietnam's new personal data protection law raises the stakes for data leaking through AI tools, driving demand for dedicated LLM firewalls.

Headquartered at 83 Science Park Drive, Singapore. V-Key took a strategic majority stake in October 2026. Holds ISO 27001:2022 and the CSA Cyber Essentials Mark; awards include the SME Cyber Security Excellence Award 2023 and Red Herring Top 100 Asia 2015. Reports 40+ enterprise customers in 8 countries and 1,300+ applications defended.