Local Voice Assistant
Qwen3-TTS voice cloning without finetuning in ~1 second on RTX 5070.
骁龙 8 Gen2 本地多模态语音 AI 助手
Full voice assistant pipeline with barge-in running entirely offline on Snapdragon GPU.
Android developers and users wanting private, offline voice interaction
Off Grid · Sherpa-ONNX
Pipeline: VAD (Silero, 16kHz) → LLM (Gemma 4 2B via LiteRT-LM on Snapdragon GPU) → TTS (sherpa-onnx + matcha-icefall-zh-baker, 22050Hz) → speaker. Barge-in interrupts TTS mid-sentence.
Demo video in the README. Code and full setup instructions on GitHub.
Qwen3-TTS voice cloning without finetuning in ~1 second on RTX 5070.
Clippy nostalgia meets screen-aware AI, but Cursor and Copilot already do this.
Runs LLM, maps, Wikipedia, and voice entirely offline on Android—matches Off Grid AI's banger pattern.
Multilingual local AI with screen sharing beats Parlor's English-only camera input.
Clean UI for Chinese name generation, but ChatGPT does this natively without a wrapper.
Nine personality modes are prompt variations wrapped in Tkinter with Groq API.