Terminal cockpit for Proxmox with AI and GitOps support
proxxx, developed by Fabriziosalmi, is a management utility for Proxmox VE and Proxmox Backup Server that centralizes cluster administration and auditability. It combines terminal-based control, command-line operations, and an embedded Model Context Protocol server to expose infrastructure state and actions to external tools. Key surface capabilities include GitOps-style state management, agentless API control, and integrated alerting for human approvals. The target audience is system administrators, DevOps engineers, and AI integrators who need auditable, automated Proxmox workflows.
What tasks can you actually use it for?
The tool focuses on operational control of Proxmox clusters: querying nodes, managing guests, and inspecting cluster status through standard APIs. It can export cluster state as stable TOML files for configuration management and supports state diff and apply operations for infrastructure-as-code workflows. For multi-profile environments it offers parallel execution and fuzzy cluster-wide search, which helps when performing batch changes across many guests or storage targets.
How secure and auditable are its operations?
Security and integrity are explicit design goals. The software implements a six-stage commit gate, SHA-256 sidecars for artifacts, sigstore keyless signatures, and zeroizes memory used for secrets. Audit logging and configuration integrity verification tools record actions for compliance purposes, supporting requirements such as NIS2 and GDPR-aligned traceability without installing agents on managed nodes.
What are the deployment and input requirements?
The distribution is a single static Rust binary for Linux and other Unix-based systems, which removes runtime dependency chains and eases installation on appliance hosts. It operates agentlessly against Proxmox VE and PBS via their APIs. The embedded MCP server is compatible with MCP-compliant hosts, allowing external assistants to connect and use exposed tool APIs for scripted queries and operations.
Does it fit into automated and AI-assisted workflows?
The built-in MCP server exposes programmatic tools so external models can query infrastructure and propose actions, while Telegram integration provides human-in-the-loop approvals for sensitive commits. GitOps-style state commands let teams keep configuration in version control, and parallel execution reduces wall-clock time for cluster-wide operations. These elements position the app for environments that require audit trails and machine-assisted orchestration rather than purely manual administration.
A practical choice for operators comfortable with code-driven infrastructure
The tool is a practical option for system administrators and DevOps teams that rely on scripted workflows and need machine-readable audit trails; it rewards teams that enforce operational discipline. Expect a learning curve for teams unfamiliar with terminal-first tools and Model Context Protocol integration, and plan governance for any automated agents interacting with infrastructure to avoid unintended changes.
