WhisperApp

Offline Download

Manually download engine binaries for use in environments without internet access

Choosing a Backend

Select the appropriate backend based on your hardware.

NVIDIA GPU → CUDA
Intel GPU / NPU → OpenVINO
AMD GPU, etc. → Vulkan
No GPU → CPU

Required Engines

WhisperApp requires the following engines to function. Download one backend per engine that matches your hardware.

whisper.cpp

Core speech recognition engine. Converts audio files and real-time audio into text.

FFmpeg

Converts audio and video file formats.

Silero VAD

Voice Activity Detection (VAD) model. Detects speech segments to improve transcription accuracy.

ModelHub

AI model download and management tool. Used to install Whisper and LLM models.

Installation

  1. 1Download the file matching your hardware from above
  2. 2Extract the ZIP into the engines folder inside the WhisperApp installation directory
  3. 3Launch WhisperApp — the placed engines will be automatically detected

Default installation path:

C:\Users\<ユーザー名>\AppData\Local\WhisperApp\engines