Buzz1.4.5

Buzz transcribes and translates audio offline on your personal computer, using OpenAI’s Whisper technology. It processes audio and video files or YouTube links, and can produce live, real-time transcriptions from a microphone. A presentation window displays live text for improved accessibility during events and presentations.

For difficult recordings, Buzz can separate speech before transcription to improve results in noisy conditions, while speaker identification labels participants in the resulting media. It supports multiple Whisper backends, CUDA acceleration for Nvidia GPUs, and Vulkan acceleration through Whisper.cpp on most GPUs, including integrated graphics. Hugging Face Whisper-type Transformer model families provide additional model options.

Transcripts can be exported as TXT, SRT, or VTT files, then reviewed in an advanced viewer with search, playback controls, speed adjustment, and keyboard shortcuts. A watch folder automatically transcribes newly added files, while the command-line interface supports scripting and automation. Its plugin system includes extensions for tasks such as AI-generated summaries and automatic transcript resizing.

Buzz suits journalists, researchers, students, accessibility staff, and anyone handling interviews, presentations, meetings, or recorded media. It runs on macOS with Apple Silicon (arm64), and its Python environment should be version 3.12; Nvidia acceleration requires CUDA support for Torch. Buzz is freeware, and version 1.4.5 is the last release that supports Intel Macs, making Apple Silicon necessary for later versions.