ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified

smallest-ai

Ultra-fast text-to-speech and speech-to-text via Smallest AI's Lightning v3.1 and Pulse models. Use when the user wants to generate speech, convert text to voice, read text aloud, create voice notes, transcribe audio to text, or clone a voice. Sub-100ms latency TTS. 64ms TTFT STT. Supports 30+ languages including Hindi and Spanish. Voices include sophia, robert, advika, vivaan, camilla, and 80+ more.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/abhishekmishragithub/smallest-ai
Or

Smallest AI — Ultra-Fast Voice Suite

Text-to-speech (sub-100ms) via Lightning v3.1 and speech-to-text (64ms TTFT) via Pulse.

Setup

  1. Get API key from https://waves.smallest.ai → click "API Key" in left panel
  2. Set SMALLEST_API_KEY in your environment:
export SMALLEST_API_KEY="your_key_here"

Defaults

  • Default female voice: sophia (American English)
  • Default male voice: robert (American English)
  • Default language: en
  • Default speed: 1.0
  • Default sample rate: 24000

Voice Selection Rules

Follow these rules to select the voice:

  1. If user explicitly names a voice (e.g. "use advika"), use that voice.
  2. If user asks for a male voice, use the configured defaultVoiceMale.
  3. If user asks for a female voice, use the configured defaultVoiceFemale.
  4. If no gender preference, use defaultVoiceFemale (sophia by default).
  5. For Hindi content: use advika (female) or vivaan (male).
  6. For Spanish content: use camilla (female) or carlos (male).
  7. For Tamil content: use anitha (female) or raju (male).

Always pass the configured defaultLanguage, defaultSpeed, and defaultSampleRate as --lang, --speed, and --rate flags unless the user overrides them.

Text-to-Speech

Generate speech audio from text using Lightning v3.1 model.

Shell (preferred — zero dependencies)

{baseDir}/scripts/tts.sh "Text to speak" --voice sophia --rate 24000 --speed 1.0 --lang en

Python (requires pip install smallestai or just requests)

python3 {baseDir}/scripts/tts.py "Text to speak" --voice sophia --speed 1.0 --lang en --out speech.wav

Voices

VoiceGenderAccentBest For
sophiaFemaleAmericanGeneral use (default)
robertMaleAmericanProfessional, reports (default)
advikaFemaleIndianHindi content, code-switch
vivaanMaleIndianBilingual English/Hindi
camillaFemaleMexican/LatinSpanish content
zaraFemaleAmericanConversational
melodyFemaleAmericanStorytelling, greetings
arjunMaleIndianEnglish/Hindi bilingual
stellaFemaleAmericanExpressive, warm

80+ more voices available. List all with: {baseDir}/scripts/voices.sh

Options

  • --voice <id>: Voice identifier (default: sophia)
  • --rate <hz>: Sample rate — 8000 | 16000 | 24000 | 44100 (default: 24000)
  • --speed <n>: Playback speed 0.5–2.0 (default: 1.0)
  • --lang <code>: Language code (default: en). See {baseDir}/references/languages.md
  • --out <path>: Output file (default: auto-named media/tts_<timestamp>.wav)

Output

Metadata

Stars4473
Views0
Updated2026-05-01
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-abhishekmishragithub-smallest-ai": {
      "enabled": true,
      "auto_update": true
    }
  }
}
Safety NoteClawKit audits metadata but not runtime behavior. Use with caution.