Enkrypt AI MCP Gateway is an open-source control plane that sits inline between AI agents and Model Context Protocol (MCP) servers to approve, modify, or block tool calls. The gateway enforces least privilege access and generates audit trails for compliance and governance workflows. The product operates as an inline MCP proxy that intercepts requests and responses, evaluating tool calls before execution and validating tool responses before they return to the agent loop. It supports three enforcement modes: modify, require approval, and block. The gateway runs guardrails directly in the request/response path to detect threats including prompt injection across text/audio/vision, tool misuse, privilege escalation, data exfiltration, jailbreaks, and response smuggling. The gateway produces policy decision logs that capture the decision, reason, and policy version applied. It maintains action traces showing tool/server, actor, environment, timestamps, and outcomes. Approval receipts document who approved what actions, when, and why. Evidence packages are generated for security reviews and governance workflows. The product is available as open source with an enterprise add-on that provides centralized policy management, tenant-aware enforcement, reporting, and support. It can be deployed as a sidecar/gateway in Kubernetes, edge, or service mesh patterns. The gateway supports multi-tenant enterprise SaaS with policies scoped by tenant/role/environment.