Voice Output Skill

Convert text to speech and speak it aloud using system TTS or browser TTS.

Setup

No additional setup required. Uses built-in system commands:

• macOS: say command (built-in)
• Linux: espeak or festival (install via apt install espeak)
• Browser: Web Speech API via Chrome DevTools Protocol

Usage

Speak text via system TTS (say command)

{baseDir}/scripts/speak.sh "Hello, the job has completed successfully!"

Speak via browser TTS (requires Chrome with CDP)

{baseDir}/scripts/speak-browser.js "Hello from the browser!"

List available voices (browser)

{baseDir}/scripts/list-voices.js

Voice Options

System voices (macOS)

List available voices:

say -v "?" 

Use a specific voice:

{baseDir}/scripts/speak.sh "Hello" --voice "Samantha"

Adjust speech rate:

{baseDir}/scripts/speak.sh "Hello" --rate 200

Browser voices

The browser TTS uses Web Speech API with available system voices. Default is usually the best available voice.

Trigger Patterns

Use this skill when:

• User asks "say", "speak", "read aloud", "text to speech"
• Job completes and you want to announce success/failure
• User wants notifications spoken rather than just displayed
• Accessibility - reading content aloud for visually impaired users
• Creating audio summaries or reports

Examples

Announce job completion

{baseDir}/scripts/speak.sh "Job complete! Processed 50 files, found 3 issues."

Read a summary

{baseDir}/scripts/speak.sh "Summary: Four files were modified, two tests added."

Browser TTS (when Chrome is available)

{baseDir}/scripts/speak-browser.js "Speaking directly through the browser!"

Notes

• System TTS (say) works on macOS out of the box
• On Linux, install espeak: sudo apt install espeak
• Browser TTS requires Chrome running with remote debugging (see browser-tools skill)
• Both methods are synchronous - the command blocks until speech completes
• For non-blocking speech, add & at the end of the command