audio.cpp
PulseAugur coverage of audio.cpp — every cluster mentioning audio.cpp across labs, papers, and developer communities, ranked by signal.
- 2026-08-01 product_launch audio.cpp released version 0.5 with new TTS and voice transfer models, alongside platform enhancements. source
- 2026-07-01 product_launch The author of audio.cpp has added support for VibeVoice 1.5B, enabling faster local text-to-speech generation. source
- 2026-07-01 product_launch The audio.cpp project released new audio models and features, including music generation, SFX generation, source separation, and long-form TTS. source
3 day(s) with sentiment data
-
Kokoro TTS integrates voice cloning via OpenVoice ToneColorConverter
A new integration allows the Kokoro text-to-speech (TTS) model to clone voices with minimal additional VRAM. By pairing Kokoro with the OpenVoice ToneColorConverter, users can achieve voice cloning capabilities without …
-
Drama3 model integrated into audio.cpp, boosting speed and reducing VRAM needs
The audio.cpp project has integrated support for the Drama3 model, significantly improving its performance. This integration allows for speeds up to five times faster than real-time on full settings, or twice real-time …
-
User experiments with voice cloning on older PC using local TTS tool
A user experimented with voice cloning on an older PC, aiming for a local text-to-speech (TTS) tool to complement Whisper. They compiled audio.cpp with Vulkan support and tested the Audio8 TTS Preview 0.6B Q8_0 model. T…
-
audio.cpp boosts performance for Higgs Audio TTS, HTDemucs, and PocketTTS
The audio.cpp project has released significant performance enhancements for various audio models. Higgs Audio TTS now requires 48% less VRAM, operating with under 6GB, while HTDemucs is 2.2x faster on GPUs and PocketTTS…
-
Reddit users compare AI audio tools: audio.cpp vs. VoiceStudio
A discussion on Reddit compares two AI-powered audio tools: audio.cpp and VoiceStudio. Audio.cpp is highlighted for its extensive model support (over 100 models) and speed, with options for local, web, and Docker deploy…
-
Yue2 model now requires only 8-9 GB VRAM
The audio.cpp project has released Yue2, a new model that requires significantly less VRAM, making it accessible for users with 8-9 GB of VRAM. This development, available in the DEV branch and on Hugging Face, offers Q…
-
audio.cpp 0.7 adds Arena UI, 62+ audio models
The audio.cpp project has released version 0.7, significantly expanding its support for audio models to 62 families and over 85 variants. This update introduces an Arena UI, allowing users to compare outputs from multip…
-
MiniMax Music 3 model generates full songs, leveraging Qwen3-8B LLM
MiniMax Music 3, a new open-weight model capable of generating complete songs up to five minutes long, has been released. This model utilizes a hybrid approach, combining an 8B Global LLM initialized from Qwen3-8B for l…
-
audio.cpp 0.5 adds DramaBox TTS, Confucius4 voice transfer, and AMD GPU support
The audio.cpp project has released version 0.5, introducing significant updates to its text-to-speech (TTS) and voice transfer capabilities. A key highlight is DramaBox, a new model designed for prompt-directed voice ac…
-
audio.cpp 0.4 adds TTS/ASR models, full GGUF support
The audio.cpp project has released version 0.4, introducing support for several new high-quality text-to-speech (TTS) and automatic speech recognition (ASR) models, including Higgs Audio v3 TTS 4B and Fish Audio S2 Pro.…
-
audio.cpp adds music, SFX, and long-form TTS models via C++/GGML
The audio.cpp project has released significant updates, introducing native C++/GGML support for several new audio models including ACE-Step, Stable Audio, HeartMuLa, RoFormer, and HTDemucs. This expansion enables music …
-
audio.cpp framework offers faster audio model inference
A new C++ inference framework called audio.cpp has been developed, built on top of ggml, to run various audio models including TTS, ASR, and voice conversion. The framework aims to consolidate multiple audio models into…