News for hackers. Stuff that matters.

settinghead

Show HN: Voxlert – SHODAN and the StarCraft Adjutant narrate my agent sessions

https://github.com/settinghead/voxlert

The problem: I run 3-10 Claude Code / Codex / Cursor sessions in parallel. When an agent finishes, I have no idea which window finished what.Voxlert gives each session a distinct character voice. Not pre-recorded sound effects — a light LLM writes an in-character phrase for each event, local TTS voice-clones the character speaking it, then audio filters (pitch, reverb) match the character's...

( Read More... | no comments | on HN )
Nice — different use case but same underlying problem. ntfy is the right call when you're away from the machine. Voxlert handles it when you're nearby but in another window: each Claude Code session gets a distinct character voice (SHODAN, StarCraft Adjutant, GLaDOS) so you hear which session needs attention without alt-tabbing.

The two actually pair well — ntfy for remote approval, Voxlert for ambient desk awareness.

https://github.com/settinghead/voxlert

On the notification side: Voxlert gives each Claude Code session a distinct character voice so you know which session needs attention without looking. SHODAN for one window, StarCraft Adjutant for another.

Pairs well with a usage bar — you'd know when a session finished and roughly what happened via audio before switching to the usage view.

https://github.com/settinghead/voxlert

Built something that takes this a step further for audio: Voxlert gives each agent session a distinct character voice (SHODAN, StarCraft Adjutant, GLaDOS, HEV Suit). When your Claude Code / Cursor session finishes or needs input, you hear a contextual phrase generated for that event — not a generic chime.

The practical thing cmux does (knowing which session needs attention without looking) is the same problem I started with. Voxlert solves it via audio identity per session rather than visual tabs.

Would love to know if there's appetite for pairing them — cmux for the terminal UI, Voxlert as an audio layer on top.

https://github.com/settinghead/voxlert