Runtime application protection platform startup Operant AI Inc. today announced the launch of AI Gatekeeper, a new product that brings end-to-end runtime artificial intelligence protection for enterprises that are deploying AI applications and agents from Kubernetes to hybrid and private clouds.

AI Gatekeeper takes Operant’s 3D Defense capabilities beyond Kubernetes with completely new defenses against rogue agents, including trust scores, agentic access controls and threat blocking for model context protocol and agentic AI nonhuman identities.

The new solution seeks to address the issue wherein enterprises are increasingly deploying sophisticated AI applications, agentic AI workflows and retrieval-augmented generation that draws on stores of data for greater model accuracy. As a result, the sprawling cloud footprint required to support such complex systems has become critically difficult to secure.

AI applications in 2025 are not only being built on cloud hyperscalers such as Amazon EKS, Fargate, Bedrock and similar services, but they’re also now expanding onto nontraditional platforms such as those from Databricks Inc., Snowflake Inc. and Salesforce Inc. Operant AI argues that the AI ecosystem and the threats that come with it are shifting closer to where the data that fuels AI actually lives, expanding security and threat exposures.

“The AI that we are now securing is a completely new beast compared to even two years ago,” said co-founder and Chief Executive Vrajesh Bhavsar. “From RAG applications to AI Agents to AI Inference systems that operate at a completely new scale, AI can’t be secured in isolation.”

AI Gatekeeper’s capabilities include comprehensive runtime defense across public, private and hybrid cloud environments. The solution extends Operant’s 3D Runtime Protection beyond Kubernetes and includes real-time catalogs of AI workloads, tools and models from providers like OpenAI, Hugging Face Inc. and Cohere Inc.

The platform also supports major large language model and data platforms while offering in-depth analytics on blocked threats, giving enterprises clear visibility into runtime threats and the security status of deployed AI systems.

AI Gatekeeper enhances cross-platform threat modeling with cohesive AI Security Graphs that map high-risk data flows. It includes out-of-the-box mappings to Open Worldwide Application Security Project Top 10 threats, such as prompt injection, data poisoning and secrets leakage, offering deep insights into affected workloads and application programming interfaces.

The platform detects supply chain risks and unauthorized AI agents using trust scores and execution boundaries. It also provides protections for MCP and AI nonhuman identies, covering both runtime and API access layers with enforced identity and access controls.

Image: Operant