FrigadeHQ
yap
Yap is a blazing-fast, on-device macOS voice dictation app that transcribes speech into any text field via a global shortcut, using Apple's macOS 26 Speech APIs with no network calls or model downloads.
audio.cpp-webui
audio.cpp is a high-performance C++ audio inference framework built on ggml that provides a unified local runtime for TTS, ASR, voice cloning, and other audio tasks, with a WebUI, CLI, server, and performance optimizations for CUDA.
Stars
245
7-day growth
No data
Forks
59
Open issues
2
License
NOASSERTION
Last updated
2026-07-23
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It replaces complex Python environments with a single portable native stack, achieving 1.8x–5x faster TTS than Python references and real-time factors as low as 0.064 for ASR, while unifying dozens of model families under one framework.
Who it is for
- Developers who want to run local audio AI models without managing multiple Python environments
- Researchers and enthusiasts exploring TTS, ASR, voice cloning, and speech processing
- Integrators building production audio pipelines that need low latency and high throughput
- Windows users looking for a straightforward local audio AI launcher with a WebUI
Use cases
- Offline text-to-speech generation with voice cloning and control
- Real-time and batch ASR for meetings, podcasts, or transcription
- Voice conversion, source separation, and diarization in a single pipeline
- Multi-step audio workflows (e.g., redubbing) using JSON pipelines
Strengths
- Performance: up to 5x faster TTS and 45–80% lower latency vs. Python references; 200x+ real-time TTS on RTX 5090
- Broad model support: 30+ model families covering TTS, ASR, VAD, diarization, source separation, codec, etc.
- Portable native stack based on ggml with CLI, server, and Gradio WebUI entry points
- Experimental pipeline system for composing multi-model workflows
Considerations
- Downstream distribution of upstream audio.cpp; not the original repository
- Requires manual build (with C++ toolchain and CMake) – no pre-built binaries for all platforms
- Some models need manual download or configuration (gated repos, missing GGUF packages), and performance on non-CUDA backends may be lower
README quick start
Usage
Description
audio.cpp with a full-task WebUI - pure C++ audio-model inference engine powered by ggml. TTS, ASR/STT, VAD, voice conversion, speaker diarization, music generation. No Python dependency.
Related repositories
Similar projects matched by category, topics, and programming language.
tetsuo-ai
voice_clone_lab
Voice Clone Lab is a local tool that fine-tunes Qwen3-TTS on a few minutes of clean speech to create a personal voice model, accessible via CLI or web UI.
Alisa0808
vox-director
An agent skill that fully automates the creation of Vox-style paper-collage explainer videos from a single topic using Atlas Cloud API and local ffmpeg.