Get protection status, backup health, and recovery readiness by asking your AI assistant, with no console switching to hunt down answers.
Make investigations recovery-aware by bringing backup health, recovery points, and threat context into the incident workflows you already run.
Spot protection gaps, unusual admin activity, and elevated-risk workloads before they turn into recovery problems.
Give AI the verified Druva backup and recovery context it needs, so recommendations and reports rest on real posture, not guesses.
Every AI interaction inherits your existing authentication and role-based access, and destructive delete operations stay blocked.
Generate operational and compliance reports by describing what you need, no query building or console work.
AI systems are only as useful as the context they reach. Without backup intelligence, they can’t tell you whether workloads are protected, where gaps create risk, or which recovery points are safe. When an incident hits, knowing what went wrong is only half the answer. Teams also need to know what they can recover. Druva MCP gives AI assistants and operations agents, like Azure SRE Agent, trusted Druva backup context inside the tools teams already use.
MCP helps make Druva’s resilience intelligence available inside the AI workflows teams already use. Instead of switching between dashboards, operators can ask questions, investigate issues, and retrieve trusted guidance through supported AI clients. In Microsoft Azure SRE Agent, for example, teams add Druva MCP as a custom MCP server so investigations can factor in backup health, protection status, and available recovery points alongside Azure operational context.
Druva MCP is designed around a Code Mode (Programmatic Tool Calling) approach that compresses the interface into five lean, multi-purpose programmatic primitives instead of hundreds of individual tool schemas. This reduces context bloat, lowers token consumption, simplifies maintenance, and guides agents toward the right workflow.
Druva MCP can support skills for user entitlements, admin and role management, endpoints, SaaS apps, enterprise workloads, cyber resilience, data governance, events, and unified reports.
Druva MCP uses a 3-legged OAuth 2.1 flow with role-based permissions, so agents act strictly on behalf of the authenticated user. If a user lacks access to a workload or action in Druva, the same access boundary natively applies through the MCP workflow.
Druva MCP runs any code or scripts in a highly restricted, ephemeral sandbox with strictly controlled outbound egress. This sandbox is completely disconnected from the Druva VPC, so AI-assisted workflows are supported without exposing broader enterprise systems or your actual backup data.