v1.6.9 v0.1.0
Z4nzu/hackingtoolAug 18, 2026 → Aug 19, 20261 day between releases
What's New
Repository adds NeMo-Speech.cpp speech stack alongside FluidVoice.
New Features
- Speech recognition with multiple AI models (ASR, Streaming, CTC)
- Speaker diarization using Streaming Sortformer v2
- Speech synthesis with MagpieTTS and NanoCodec
- Text and speech translation via Riva Translate 4B Instruct v2
- Unified CLI with managed model discovery, downloads, and caching
Improvements
- OpenAI-compatible HTTP API for transcription and speech subsets
- Native C API and CMake package for application integration
- Support for CPU, CUDA, Vulkan, and Metal backends
- Subtitles, VAD, endpointing, punctuation, and capitalization features
- Realtime WebSocket transcription and playground
Upgrade Notes
The repository now includes the NeMo-Speech.cpp module alongside FluidVoice. Note that v0.1.0 is an early release and API interfaces may evolve. Ensure your environment supports the required backends (CUDA, Metal, Vulkan, or CPU).
The Two Releases
FluidVoice v1.6.9
Aug 18, 2026· 2 months ago
This release focuses on stability improvements for microphones and audio transcription, alongside expanded formatting capabilities for the Custom Dictionary.
Full release notesv0.1.0
Aug 19, 2026· 2 months ago
v0.1.0 is the inaugural release of NeMo-Speech.cpp, introducing a comprehensive suite of speech processing tools including recognition, diarization, synthesis, and translation, backed by multiple AI models.
Full release notes