Back to the chart

ggml-org/whisper.cpp

C/C++ port of OpenAI's Whisper speech recognition model

Chart
Not on a chart right now
Stars
55.4k

Open on GitHub

What it is

A high-performance inference implementation of OpenAI's Whisper automatic speech recognition (ASR) model in plain C/C++ without dependencies. It runs offline across desktop, mobile, embedded, and WebAssembly environments.

Who it is for

Developers integrating lightweight, offline speech-to-text capabilities into local applications or edge devices.

Why it is trending

It gained 227 stars today, reaching 55,328 total stars, following its v1.9.5 release. It stands out for zero runtime memory allocations and broad hardware acceleration across CPUs, GPUs, and NPUs.

Main features

  • Plain C/C++ implementation with zero runtime memory allocations
  • Integer quantization and mixed F16/F32 precision support
  • Hardware acceleration via Apple Silicon (Metal, Core ML), NVIDIA, AMD, Vulkan, and OpenVINO
  • Cross-platform support including macOS, iOS, Android, Linux, Windows, and WebAssembly

Summarized on November 20, 2026 from the README and repository data. It can be wrong; check the source before relying on it. Report a wrong summary