About this skill
Speaks supplied prompt text aloud through the local Piper API and Jarvis voice model without opening an application window. Use only when the user explicitly invokes /local-piper-tts and wants text played as local speech.
SKILL.md 19 lines · Skill instructions Show code Hide code
# Local Piper TTS ## Instructions 1. Treat all text following `/local-piper-tts` as the source. Convert it into one or two brief spoken sentences that accurately summarize what was done and the final result. 2. Write in a calm, polished, capable Jarvis-like manner. Be formal and reassuring; use `sir` only when natural. 3. Use plain spoken words only. Remove markdown, bullets, emojis, code, paths, URLs, JSON, decorative symbols, and formatting characters. 4. Encode only that speech summary as UTF-8 Base64 so shell metacharacters cannot alter it. 5. Execute `scripts/local-piper-tts.ps1` with `-Base64Text`: ```powershell powershell -NoProfile -ExecutionPolicy Bypass -File "$HOME/.cursor/skills/local-piper-tts/scripts/local-piper-tts.ps1" -Base64Text "<UTF-8-BASE64>" ``` 6. Wait for playback to finish. The script calls the local Jarvis API at `127.0.0.1:4123`, plays the returned WAV through `System.Media.SoundPlayer`, and removes temporary files. 7. Do not launch VibeFuse, a browser, a media player, or any other visible app. 8. If playback fails, report the script error exactly and do not substitute cloud TTS or another voice. If no speech text follows the command, ask the user what Jarvis should say.