Buzz
Buzz transcribes and translates audio offline on your personal computer. Powered by OpenAI's Whisper.
open-sourcevoice-agents
21.8k
Stars
+534
Stars/month
41
Commits (90d)
1
Releases (6m)
Star Growth
+3.4k (18.4%)
Overview
Buzz 是一款基于 OpenAI Whisper 技术的本地音频转录和翻译工具，支持离线处理以保护隐私。该工具可以转录音频和视频文件、YouTube 链接，还提供实时麦克风转录功能。Buzz 支持语音分离技术以提高噪音环境下的转录准确性，具备说话人识别功能，并支持多种 Whisper 后端（包括 CUDA、Apple Silicon 和 Vulkan 加速）。工具提供高级转录查看器，具备搜索、播放控制和速度调整功能，支持导出为 TXT、SRT 和 VTT 格式。监视文件夹功能可自动转录新文件，命令行接口支持脚本化和自动化操作。作为一款跨平台工具，Buzz 在 GitHub 上获得了 18,429 个星标，证明了其在音频转录领域的受欢迎程度。
Deep Analysis
Key Differentiator
vs Whisper CLI: full GUI with live transcription, speaker ID, and watch folders; vs cloud transcription (AssemblyAI/Deepgram): completely offline with zero data leaving the device
⚡ Capabilities
- • Offline audio transcription with Whisper
- • Live real-time microphone transcription
- • Video and YouTube link transcription
- • Speaker identification (diarization)
- • Speech separation for noisy audio
- • Multiple Whisper backend support
- • Export to TXT/SRT/VTT formats
- • Watch folder for automatic transcription
🔗 Integrations
OpenAI WhisperWhisper.cppCUDA (Nvidia GPU)Apple SiliconVulkan GPU acceleration
✓ Best For
- ✓ Offline audio/video transcription with privacy
- ✓ Live presentation captioning
- ✓ Batch transcription of media files
✗ Not Ideal For
- ✗ Cloud-scale transcription services
- ✗ Real-time low-latency streaming transcription APIs
Languages
Python
Deployment
macOS DMGWindows installerLinux Flatpak/Snappip install (PyPI)
Pricing Detail
Free: Fully open-source MIT license
Paid: N/A - free
⚠ Known Limitations
- ⚠ Transcription speed depends on hardware
- ⚠ Large models require significant RAM/VRAM
- ⚠ Speaker identification accuracy varies
- ⚠ No cloud API - desktop application only
Pros
- + 完全离线处理，保护用户隐私，无需将音频数据上传到云端
- + 支持多平台和多种 GPU 加速（CUDA、Apple Silicon、Vulkan），提供优化的性能
- + 功能全面，包括实时转录、说话人识别、语音分离和多种导出格式
Cons
- - Windows 版本未签名，安装时会出现安全警告
- - PyPI 安装需要特定的 Python 3.12 环境和 ffmpeg 依赖
- - 高质量转录可能需要较强的硬件配置以支持 GPU 加速
Use Cases
- • 转录采访、会议或播客内容，生成可搜索的文本记录
- • 为视频内容创建字幕文件（SRT、VTT 格式），提高内容可访问性
- • 在演示、讲座或会议期间提供实时字幕，支持无障碍访问
Getting Started
1. 从 SourceForge 下载对应平台的安装包（macOS 下载 .dmg，Windows 下载安装文件，Linux 使用 Flatpak 或 Snap）并完成安装；2. 启动 Buzz 应用程序，首次使用时可能需要下载 Whisper 模型；3. 选择音频或视频文件进行转录，或点击麦克风按钮开始实时转录，转录完成后可导出为所需格式
Alternatives
w
whisperX
WhisperX: Automatic Speech Recognition with Word-level Timestamps (& Diarization)
W
WhisperS2T
An Optimized Speech-to-Text Pipeline for the Whisper Model Supporting Multiple Inference Engine
Compare Buzz
Maintain Buzz?
Show your live rank in your README, or put Buzz in front of every visitor to AgentoolRank.