← All projects

xbxd/splashbar

SplashBar

A Mac menu bar app for checking and controlling my local AI setup. I wanted the useful status and a few everyday controls close at hand, without opening a terminal.

  • experiments
  • macos

Project details

Primary language: Python

Last commit:

Language mix

  • Python 76%
  • Swift 24%

Weekly commits

Commit counts by UTC week start for the last 52 weeks.

74 commits total; busiest week: 2026-10-04 (41 commits).

README

SplashBar

SplashBar is a small macOS menu bar app I made to keep an eye on, and control, my local AI setup.

Status

This is an experiment now, not a tool I actively use or maintain. I still like having it here as a snapshot of a project I built.

What's in here

  • SplashBar/ is the native menu bar app, built with SwiftUI and AppKit.
  • SplashBar/Sources/SplashBar/ contains the status item, menu, and shortcut UI.
  • SplashBar/Sources/SplashBarCore/ holds app state, backend communication, dependency presentation, and shortcut logic.
  • SplashBar/Sources/SplashBarCoreChecks/ has small executable checks for core presentation and shortcut behavior.
  • SplashBar/preview/ contains helpers for previewing the app's presentation.
  • scripts/ contains the Python backend and local service controls. The app uses these to read status and offer actions from its menu.
  • swiftbar/ is a SwiftBar plugin version, along with the small robot icons it uses.
  • tests/ has offline Python tests for the scripts and plugin.
  • docs/ has project notes and design history.
  • skills/ has a project-specific guide for working with the local setup.

How I use it

I built SplashBar as a quick way to see whether the local service is up, starting, sleeping, or busy, without leaving the menu bar. From the menu, it can refresh its view, start or stop the service, and offer a few confirmed settings and maintenance actions.

The native app is written in Swift 6 and targets macOS 14.4 or later. The companion controls and SwiftBar plugin are Python.

Notes

SplashBar is for a local setup. It does not expose the service to the network or connect it to an AI client.