eviscerations/whisper-windows-mcp
🔌 MCP Servereviscerations
Windows-native local transcription server using whisper.cpp with Vulkan GPU acceleration for private, offline audio processing.
The whisper-windows-mcp server bridges the gap between local Windows hardware and AI-powered transcription. Built specifically for the Windows ecosystem, it eliminates the dependency on heavy Python runtimes or external cloud-based APIs, ensuring complete data privacy and offline functionality. The core engine utilizes whisper.cpp, optimized with Vulkan GPU acceleration to deliver fast, efficient inference on consumer-grade hardware. Key features include comprehensive model management for downloading and switching between different Whisper model sizes, support for multiple languages, and a background job handler that allows for asynchronous batch processing of audio and video files. By implementing the Model Context Protocol, this tool allows AI assistants and agents to seamlessly request transcriptions from local files, providing a standardized way to handle media input in local-first AI architectures.
💡Highlights
- ├─Vulkan-accelerated whisper.cpp
- ├─Zero cloud API dependency
- └─Native Windows background jobs
🎯For
- ├─AI Application Developers
- └─Privacy-focused Engineers