| name | heyboy-voice-assistant |
| description | Install, configure, diagnose, and operate the heyboy-voice-assistant CLI voice loop. Use for requests to set up "hey boy" wake-word workflows, fix missing heyboy command/PATH issues, switch backends (OpenClaw API, Codex CLI, Claude CLI, generic CLI), or manage always-on background mode on macOS. Includes best-effort Windows setup notes. |
heyboy-voice-assistant
Use this workflow to run HeyBoy as a CLI-first voice assistant.
1) Detect platform and command availability
- Run
command -v heyboy on macOS/Linux.
- Run
Get-Command heyboy on Windows PowerShell.
- If command exists, skip install and move to setup.
2) Install when heyboy is missing
macOS (tested)
curl -fsSL https://raw.githubusercontent.com/EthanSK/heyboy-voice-assistant/main/scripts/install.sh | bash
- If shell cannot find
heyboy, apply PATH update guidance in references/macos.md.
Windows PowerShell (best-effort, untested)
iwr https://raw.githubusercontent.com/EthanSK/heyboy-voice-assistant/main/scripts/install.ps1 -UseBasicParsing | iex
- If shell cannot find
heyboy, apply PATH update guidance in references/windows.md and restart terminal.
3) Configure backend
Run exactly one backend setup:
heyboy setup openclaw --api-key "YOUR_TOKEN"
heyboy setup codex
heyboy setup claude
heyboy setup generic --command "ollama run llama3.2"
4) Validate environment
Run:
heyboy doctor
If doctor fails because dependency/model is missing, run:
heyboy install
Then re-run doctor.
5) Start assistant
Foreground loop:
heyboy run
macOS background app mode (LaunchAgent):
heyboy app install
heyboy app start
heyboy app status
Stop/uninstall:
heyboy app stop
heyboy app uninstall
6) Platform notes
- Read
references/macos.md for tested commands and log locations.
- Read
references/windows.md for best-effort setup scripts and current limitations.
- Read
references/ui-validation.md before making claims about desktop UI readiness.