Wardn AI is a self-hosted control plane for governed AI tool access. It lets teams curate MCP servers, protect credentials, run agents, and trace every action in one place.
It is built for platform engineering, security, and AI teams that need approved tools, reusable agents, and human approval routes. Wardn keeps tool access, policy decisions, token use, and costs in the same execution path, with support for local or OIDC auth, OpenBao secret backends, and Kubernetes-oriented runtime isolation.
Wardn can run agents through chat, schedules, Slack, Telegram, or WhatsApp, and it records tool calls, approvals, failures, and usage for each run.
Follow the project
→Built by