Download

Get Kuma for your computer.

Free, open source, and fully offline. Builds for macOS, Windows and Linux are published on GitHub. Pick the installer for your system on the releases page.

Pick your platform

macOS

Windows

Linux

One requirement: Kuma uses ffmpeg to read audio and video, so it needs to be installed and on your PATH. Most Linux systems and many Macs already have it; otherwise install it with Homebrew (brew install ffmpeg) or your package manager. Bundling ffmpeg so this is not needed is on the roadmap.

First run

When you open Kuma, go to the Models panel and download a model, or open African languages to pick the right one for your language. Then drop in an audio or video file, or record from your microphone, and it transcribes on your device. The models guide explains how to add languages that stock Whisper does not cover.

Build from source

Prefer to build it yourself, or want a specific GPU backend? You need Rust, Node, cmake and ffmpeg:

git clone https://github.com/geniuskidkanyi/kuma
cd kuma
npm install
npm run tauri build

To enable a specific GPU backend on Windows or Linux, pass a feature, for example npm run tauri build -- --features vulkan (also cuda, rocm or sycl). macOS enables Metal automatically. Installers are written to src-tauri/target/release/bundle/.


Hit a problem installing? Open an issue and include your operating system.