LocalKin
Complete Voice AI Platform
for Local Development
Run Speech-to-Text, Text-to-Speech, and LLM integration locally with an Ollama-inspired interface.
Available for macOS, Linux, and Windows
Key Features
Speech-to-Text
Multiple STT engines including whisper.cpp (50x faster), faster-whisper with VAD, and OpenAI Whisper
whisper.cppfaster-whisperWhisper
Text-to-Speech
Multiple TTS engines with voice selection: Native OS TTS, Kokoro, SpeechT5, Bark, and XTTS for voice cloning
KokoroXTTSSpeechT5
LLM Integration
Voice-based conversational AI with Ollama integration for intelligent responses and streaming support
OllamaStreamingContext
Quick Start
# Install with uv (recommended)
1
uv pip install localkin-service-audio# Transcribe audio
2
kin audio transcribe audio.wav# Text-to-speech
3
kin audio tts "Hello world" --model kokoro-82m# Voice AI with LLM
4
kin audio listen --llm ollama --tts --stream# Start API server
5
kin audio run kokoro-82m --port 8001