Home Assistant Integration to install openWakeWord´s from fwartner/home-assistant-wakewords-collection

A voice assistant with local LLM as a backend

Train custom wake word models with openWakeWord. A granular 13-step pipeline with compatibility patches for torchaudio 2.10+, Piper TTS, and speechbrain. Generates tiny ONNX models (~200 KB) for real-time keyword detection — like building your own "Hey Siri" trigger. WSL2/Linux + CUDA required.

Trigger-Talk is an offline-capable hotword detection framework that passively listens for custom wake phrases to trigger speech recognition or automation workflows.

Custom OpenWakeWord training pipeline for generating wake word models compatible with HAwake Android

Small browser-first wrapper around the OpenWakeWord models using onnxruntime-web. It exposes a WakeWordEngine class you can drop into a React app to listen for wake words like hey_jarvis directly in Chrome, no native layer required.

Offline AI voice assistant with semantic memory, wake-word detection, local LLM inference, streaming TTS, and modular tool-agent architecture.

A cpp runtime + c shared library for openwakeword detection

Hands-free voice input for Claude Code on Linux

Angela Anthropomorphized - 3d AI Voice Assistant

A Dockerized, GPU-first trainer for openWakeWord custom models with a Web UI. Generates synthetic positives with Piper (and optional ElevenLabs voices), pulls real-world augmentation corpora (MIT IR Survey, MUSAN, FSD50K, Common Voice), trains a small classifier head on top of Google's frozen speech-embedding model.

Bulletproof openWakeWord trainer Colab notebook (2026 edition). Run-all-and-walk-away. Train your own wake word in 75-90 min on Colab Pro.

Meet Vi, a modular, voice-activated AI assistant built in Python. It utilizes an intent router to process speech-to-text input and dynamically dispatch commands to specialized, locally hosted Hugging Face models (Summarization, Sentiment Analysis, and Conversational QA) based on context. A lightweight exploration of agentic AI workflows.

Wake word Gemma (gem-muh)

Local-first modular voice assistant with Rasa, FastAPI, and Whisper — designed for privacy, extensibility, and clean system architecture.

A wrapper for the openWakeWord library

Open source voice assistant inspired by Jarvis. Powered by Claude with real-time STT, streaming TTS, wake-word detection, and a live web dashboard. Turkish-first, self-hosted.

Holographic AI assistant with a real-time voice pipeline, multimodal LLM routing, and Lightning Network payments — all behind a Three.js HUD with custom GLSL shaders. Ships as a standalone .exe.

AgeniusNote: voice-first scratchpad. Local STT (Whisper), wake words (openWakeWord), optional AI parse via Ollama/OpenAI/Anthropic. PySide6 desktop. MIT.

Just say "Hey Echo." A local-only Windows tray app that turns GitHub Copilot into a voice assistant — wake word detection, multi-turn conversations, project knowledge bases, and autonomous routines like morning standup prep. All speech processing runs locally; interrupt anytime with your voice or a triple-tap of Caps Lock.