← Hongjun Ma – Technical Projects
An open-source MCP server for HPE compute: ask your AI assistant about your HPE ProLiant servers through HPE iLO (Redfish), HPE Compute Ops Management (GreenLake) and HPE OneView / Synergy.
GitHub repository PyPI package Tool reference Changelog
proliant-mcp is a Model Context Protocol (MCP) server. It gives AI assistants such as GitHub Copilot, Claude, Codex and opencode read-only access to your HPE server management tools, so you can ask questions in plain English instead of clicking through iLO, GreenLake or OneView consoles.
Ask things like:
Option 1 – ask your AI agent. Paste this into GitHub Copilot, Claude, Codex or opencode:
Install the proliant-mcp MCP server for me. Follow the instructions at https://raw.githubusercontent.com/hjma29/proliant-mcp/main/AGENTS.md (install uv first if it is missing), then tell me when to restart you.
Option 2 – manual. Install uv, then add this to your AI client's MCP configuration and restart the client:
{"servers": {"proliant": {"type": "stdio", "command": "uvx", "args": ["proliant-mcp@latest"]}}}
In Claude Desktop use mcpServers instead of servers. Then just ask your assistant to "sign in to HPE GreenLake", "add my iLO" or "add my OneView". Passwords are typed on a page on your own computer, never in the chat.
| Tool | What it returns |
|---|---|
get_server_summary / get_fleet_summary | Model, serial, power and POST state, health, BIOS and iLO version, CPU/memory, running OS and boot order – for one or all iLOs |
get_health, get_firmware_inventory | Per-subsystem health, fan/PSU redundancy, every firmware component |
get_processors, get_memory, get_storage, get_network_adapters | CPUs and GPUs, DIMMs, RAID controllers and drives, NICs with LLDP switch neighbors |
get_event_log, get_fleet_event_summary, get_ilo_logins | IML, IEL and Security Log filtering and summaries, iLO login audit |
| Tool | What it returns |
|---|---|
com_get_overview | Fleet health, power, connection and subscription report with servers needing attention |
com_list_servers, com_get_server | Server inventory and per-server detail, firmware and alerts |
com_hardware_report, com_list_nics, com_list_gpus | Whole-fleet hardware, NIC/HBA and GPU reports in one call |
com_list_alerts, com_list_jobs, com_list_groups, com_list_firmware_bundles | Alerts, firmware-update jobs, group compliance, SPP bundles |
| Tool | What it returns |
|---|---|
oneview_get_appliance, oneview_list_alerts | Appliance health and active alerts with corrective actions |
oneview_list_server_hardware, oneview_list_server_profiles | Servers, profiles, templates and firmware compliance |
oneview_list_enclosures, oneview_list_interconnects, oneview_list_uplink_sets | Synergy frames, Virtual Connect modules and uplink status with LLDP neighbors |
See the full tool reference on GitHub.
All server tools are read-only. Credentials stay on your computer and are never sent to the AI model. The setup tools only edit a local inventory file and never take a password as an argument.