Enkrypt AI

Enkrypt AI is an enterprise AI security and compliance platform focused on protecting AI models, agents, and multimodal systems in production. Its core mechanism is red teaming: automated, adversarial testing that probes deployed models and agents for weaknesses like prompt injection, jailbreaking, data leakage, and unsafe or non-compliant outputs. Failures found during red teaming feed into Agent Guardrails, policy-based controls enforced in real time to block the same issues at runtime. The platform also includes an Agent Policy Engine that ingests governance documents and regulations and converts them into enforceable controls traceable back to the source clause, an AI Data Risk Audit for models and embeddings, an MCP Gateway and Scanner for securing Model Context Protocol servers and tool calls, an AI Asset Scanner for inventorying AI deployments, and a public LLM Safety Leaderboard benchmarking over 200 models. Enkrypt AI serves security, compliance, and engineering teams at enterprises in regulated industries such as finance, healthcare, insurance, and government; named customers include Salesforce, NetApp, Skyhigh Security, and AI21 Labs. It was founded by Yale PhD researchers in 2022 and offers SDKs and API documentation for integrating its guardrails into existing AI pipelines. Detailed pricing is not published; the site directs prospective customers to request a demo or trial.