v2.10 v0.1.0
caddyserver/caddyAug 19, 2026 → Aug 19, 20260 days between releases
What's New
NeMo-Speech.cpp v0.1.0 introduces speech recognition, synthesis, and translation capabilities.
New Features
- Speech recognition using Nemotron 3.5 ASR, Parakeet TDT, and CTC
- Speaker diarization with Streaming Sortformer v2
- Speech synthesis with MagpieTTS and NanoCodec
- Text and speech translation with Riva Translate 4B Instruct v2
- Local HTTP API with OpenAI-compatible transcription and WebSocket support
Improvements
- Unified nemo-speech CLI with managed model discovery
- Support for CPU, CUDA, Vulkan, and Metal backends
- Native C API and shared libraries for integration
- Subtitles, VAD, endpointing, and punctuation support
Upgrade Notes
NeMo-Speech.cpp is a new project with no upgrade path from Caddy v2.10. This is an early release where API interfaces may evolve.
The Two Releases
v2.10
Aug 19, 2026· 2 months ago
This release features a major overhaul of the internal browser, improvements to Codex and Responses API handling, the bundling of the Agent Coordination Protocol (ACP), and unified parallel calls for subordinates.
Full release notesv0.1.0
Aug 19, 2026· 2 months ago
This is the first versioned release of NeMo-Speech.cpp, introducing a comprehensive suite of speech processing capabilities including recognition, synthesis, and translation with support for multiple backends and platforms.
Full release notes