Higangssh/homebutler

🏠 Tells you what changed on your server — only what's worth telling. Single Go binary, no database, nothing running on the machines it watches. MCP server built in.

What it solves

HomeButler addresses the "messy middle" of homelab maintenance, where servers become difficult to manage as they drift over time. It solves the problem of not knowing what has changed on your server since the last time you checked, helping you identify why services restart, whether ports are occupied by the wrong processes, and if backups are actually restorable.

How it works

It is a Go binary that functions as an operations toolkit without requiring a persistent daemon or database. It takes snapshots of your server's state (containers, ports, processes) and compares them against previous baselines. It provides a structured, butler-style report of changes (e.g., "new," "gone," "replaced") and includes a "doctor" command for pre-flight checks and a "drill" feature to verify that backups can successfully boot and pass health checks.

Who it’s for

This is for self-hosters and homelab enthusiasts who want a clear, structured interface to monitor their servers. It is also designed to be used by AI agents or automated scripts, as it provides machine-readable JSON output for all its diagnostic and monitoring functions.

Highlights

  • Change Detection: Automatically identifies and categorizes changes like container recreation, image updates, or port ownership shifts.
  • Backup Verification: Includes a drill command that boots backups in isolated containers to confirm they are functional.
  • Agent-Friendly: Designed to be used by AI agents via MCP (Model Context Protocol) or CLI, providing structured data without needing full SSH shell access.
  • No Daemon Required: Operates as a standalone tool that runs over SSH, keeping your server footprint minimal.
  • Crash Analysis: Monitors Docker, systemd, and PM2 services to capture logs and analyze crash causes like OOM (Out of Memory) or segmentation faults.

Related

  • Project
    Danceiny/gotry

    GoTry is an AI travel agent that separates conversational language processing from deterministic code-based calculation to provide reliable, evidence-backed travel itineraries.

  • Project
    ucsandman/DashClaw

    A safety and governance layer for unattended AI agents that intercepts destructive tool calls and requires human approval via a signed decision ledger.

  • Project
    SakanaAI/mass

    MASS is a research framework that enables recursive self-improvement in AI models by alternating between optimizing multi-agent workflows and using the resulting data to fine-tune the model.

  • Project
    Signet-AI/signetai

    Signet AI is an open‑source service that automatically extracts and structures information from files, chats and web sources into a searchable knowledge graph. Connected AI agents (Claude Code, OpenCode, Gemini CLI, etc.) can query this graph for relevant context before generating responses, preserving memory across model switches. It offers real‑time sync with Obsidian, GitHub, Notion, Discord, a background “dreaming” pipeline, a web/desktop dashboard, encrypted secret storage, and optional telemetry. Installable via a one‑liner script, npm/Bun, or Docker, it runs locally on macOS, Linux and Windows.