HiveTrace Deploys AI Firewall in Cloud.ru to Protect Corporate GenAI Applications

HiveTrace, a developer of artificial intelligence protection solutions, has deployed its AI Firewall in the Cloud.ru cloud. This move accelerates implementation for corporate customers and increases the level of protection for GenAI applications against the most common attacks on large language models. AI Firewall protects against key information security threats when using AI, including prompt attacks, data and model poisoning, leakage of system instructions, excessive agency, and other LLM vulnerabilities. For example, it prevents attackers from forcing the model to perform unsafe actions through malicious queries and from obtaining confidential company data via the LLM. Unlike protection tools at the level of individual models, the solution works at the application level. In addition to the migration, Cloud.ru and HiveTrace agreed to include the HiveTracePro model in the Evolution Foundation Models service. Clients of Cloud.ru who already use language models on the platform can connect HiveTracePro as an additional level of protection for GenAI applications. The model complements Guardrails Filter, Cloud.ru's own tool for masking sensitive data when working with large language models. The clients of HiveTrace work in various industries, so infrastructure requirements differ from project to project. The company needed a unified platform that meets all requirements for working with personal data, GIS, critical information infrastructure, and financial systems, and also supports Kubernetes with GPU. For hosting AI Firewall, the company chose Cloud.ru Evolution. The solution was deployed on the basis of Evolution Managed Kubernetes with GPU. Cloud.ru's managed services turned out to be compatible with HiveTrace's existing components, so no additional refinement was required.