characteristic-voice — Codex skill: what it does, how to install
Adds emotion, pauses and conversational delivery to synthesized speech. Use it for voice messages with a chosen mood or a voice reference.
npx skills add NoizAI/skills --skill characteristic-voice -a claude-codeWhat characteristic-voice does
This skill shapes TTS output with presets for morning, bedtime, comfort, celebration and casual chat. It can adjust emotional tone, pace and small vocal fillers so a line feels less mechanical. The included speak.sh script wraps the tts skill.
You can override emotion settings, select a voice and choose an output format. The --ref-audio option supplies a voice reference for cloning, using a short WAV clip. Reusing recordings from other people or copyrighted media can raise rights issues.
speak.sh always requires curl and python3. The Noiz backend needs NOIZ_API_KEY and sends spoken text plus reference audio off-device, so choose --backend kokoro when local processing matters. Kokoro requires kokoro-tts; reference download and trimming use yt-dlp and ffmpeg.
What it helps with
- generate a gentle bedtime voice message
- set a comforting emotional tone for speech
- adjust emotion and speech speed
- clone a voice from reference audio
- run speech generation locally with Kokoro
Example prompts for the agent
- Create a bedtime audio message with the
goodnightpreset. - Read this line with tenderness using
--emoand speed0.75. - Generate a voice message using
skills/characteristic-voice/character.wavwith--ref-audio.
How to install characteristic-voice
1Skills CLI (Vercel)
npx skills add NoizAI/skills --skill characteristic-voice -a claude-code2Manually: put the files in place
~/.claude/skills/characteristic-voice/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/characteristic-voice/
1Skills CLI (Vercel)
npx skills add NoizAI/skills --skill characteristic-voice -a codex2Manually: put the files in place
.agents/skills/characteristic-voice/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add NoizAI/skills --skill characteristic-voice -a cursor2Manually: put the files in place
.cursor/skills/characteristic-voice/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install NoizAI/skills characteristic-voice2Manually: put the files in place
.github/skills/characteristic-voice/SKILL.md1Skills CLI (Vercel)
npx skills add NoizAI/skills --skill characteristic-voice -a gemini-cli2Platform CLI
gemini skills install https://github.com/NoizAI/skills --path skills/characteristic-voiceManually: put the files in place
.devin/skills/characteristic-voice/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add NoizAI/skills --skill characteristic-voice -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about characteristic-voice
How do I install characteristic-voice in Codex?
Run npx skills add NoizAI/skills --skill characteristic-voice -a codex. The installed skill file is at .agents/skills/characteristic-voice/SKILL.md.
What does characteristic-voice do?
It prepares expressive TTS speech with presets, pacing, fillers and emotion controls. Its speak.sh script produces audio through the tts skill.
What is required for the local Kokoro backend?
speak.sh requires curl and python3, while Kokoro also needs kokoro-tts. Kokoro processes speech locally.