Smallest AI — OpenClaw Voice Skill

The fastest TTS/STT skill for OpenClaw. Sub-100ms text-to-speech via Lightning v3.1. 64ms speech-to-text via Pulse.

"My OpenClaw now speaks faster than I can think" — what you'll say after installing this

Why This Exists

Provider	Latency (TTFB)	Price	Voice Cloning
Smallest AI	~100ms	Free tier + $5/mo	5 sec of audio
ElevenLabs	~300ms	$5/mo (30min)	30 sec of audio
OpenAI TTS	~200ms	Pay per token	Not available
Kokoro (Local)	Varies	Free (GPU needed)	Not available

Smallest AI is faster, cheaper, and needs 6x less audio for voice cloning.

Quick Install

# Option 1: ClawHub (recommended)
clawhub install smallest-ai

# Option 2: Manual
git clone https://github.com/smallest-inc/smallest-ai-openclaw.git
cp -r smallest-ai-openclaw ~/.openclaw/skills/smallest-ai

# Set your API key
export SMALLEST_API_KEY="your_key_here"
# Get one free at https://waves.smallest.ai

Then restart the gateway: openclaw gateway stop && openclaw gateway start

What You Get

Text-to-Speech (Lightning v3.1)

"Say good morning in a male voice"
"Read my latest email summary aloud"
"Generate a voice note saying the meeting is at 3pm"

Speech-to-Text (Pulse)

"Transcribe this voice note" [attach audio]
"What did they say in this recording? Include speaker labels"
"Convert this meeting audio to text with timestamps"

Multilingual

"Say 'नमस्ते, कैसे हैं आप?' in Hindi"
"Read this in Spanish: Hola, buenos días"
"Say bonjour le monde in French"

Voice Note Loop

Send a WhatsApp voice note → Pulse transcribes it → agent processes → replies as a voice note via Lightning. Full voice-in, voice-out.

Voices

Voice	Gender	Accent	Best For
sophia	Female	American	General use (default)
robert	Male	American	Professional, reports (default)
advika	Female	Indian	Hindi, code-switching
vivaan	Male	Indian	Bilingual English/Hindi
camilla	Female	Mexican/Latin	Spanish content
ella	Female	American	Conversational
mia	Female	American	Storytelling
arjun	Male	Indian	English/Hindi bilingual
vanessa	Female	American	Expressive, warm

80+ more voices available — the agent auto-selects based on language and gender preference.

File Structure

smallest-ai/
├── SKILL.md                 # Skill definition (required)
├── scripts/
│   ├── tts.sh               # TTS via curl (zero deps)
│   ├── tts.py               # TTS via Python (SDK + fallback)
│   ├── stt.sh               # STT via curl
│   ├── stt.py               # STT via Python
│   └── voices.sh            # List available voices
├── references/
│   ├── voices.md            # Voice catalog
│   ├── languages.md         # 30+ supported languages
│   └── api-reference.md     # API quick reference
└── README.md                # This file

Requirements

• OpenClaw (any version)
• curl (included on macOS/Linux)
• python3 (optional, for advanced features)
• SMALLEST_API_KEY environment variable

Optional: pip install smallestai for the official SDK with async support and streaming.

Examples

Voice Note Loop (WhatsApp)

Send a voice note saying "summarize my day" →
Agent transcribes → processes → replies with voice note

Morning Briefing Agent

"Every morning at 7am, check my calendar and email,
 summarize the day ahead, and read it to me on WhatsApp
 using the robert voice"

Hindi/English Code-Switching

"Hey, mujhe aaj ke weather ka summary do in advika's voice"

Meeting Transcription

"Transcribe this meeting recording, identify who said what,
 and create a summary with action items"

Pricing

Plan	TTS	STT	Voice Clones	Cost
Free	30 min/mo	Limited	0	$0
Basic	3 hrs/mo	Included	1	$5/mo
Premium	24 hrs/mo	Included	2	$29/mo

Links

• Smallest AI — Main site
• Waves Console — Get API key
• API Docs — Full documentation
• Python SDK — Official SDK
• OpenClaw — OpenClaw main site
• ClawHub — Skills marketplace

Contributing

PRs welcome! If you add a new feature or voice integration, please include:

1. Updated SKILL.md instructions
2. Tests or usage examples
3. Updated references if API surface changes

License

MIT — use it, fork it, ship it.

---

Built with ⚡ by Abhishek at Smallest AI