TOOLORA / Local reading edition
whisper.cpp
Run Whisper transcription locally in C/C++, with CPU, Metal, CUDA, Vulkan and quantization support for desktop, mobile and browser integration.
Overview
whisper.cpp is ggml-org’s Whisper inference implementation, with a C-style API and command-line examples that turn audio into transcripts and timing information. Backends include CPU, Apple Metal, CUDA and Vulkan, alongside quantization and VAD. Examples cover iOS, Android and WebAssembly integration. After the required model files are prepared, suitable builds can perform local inference; available acceleration and output options depend on the platform and build configuration.
Official sources and references
- Official websitehttps://github.com/ggml-org/whisper.cpp
Descriptions summarize public information; consult official websites for features, pricing and regional availability.