Four on-device TTS models on one CPU core: LinnetTTS, Piper, Kokoro, pocket-ttsFour open-source CPU TTS models compared: RTF 5.8% to 51.7%, 187 MB to 1.2 GB resident, voice cloning, streaming, and license trade-offs across models6 min read·voxrt.hashnode.dev6d ago
Building a fully offline voice assistant on Raspberry Pi 4Step-by-step guide to wake-word, keyword spotting, and streaming ASR running entirely on-device on $35 hardware. Code in Python, Node.js, Go, and C.14 min read·voxrt.hashnode.devOct 30
On-device streaming ASR in 2026: five approaches and when to pick whatComparing VoxRT, Picovoice Cheetah, sherpa-onnx, Vosk, and whisper.cpp with a decision guide.10 min read·voxrt.hashnode.dev2Oct 23