• English
      • ManySpeech
      Lightweight Local
      Speech Processing Toolkit

      Native offline recognition & real-time streaming recognition, no cloud dependency, zero-config startup

      Quick StartGitHub

      🎯 Dual-Core Driven

      Offline audio recognition + microphone real-time streaming recognition, covering all scenarios

      🤖 Smart Model Management

      Automatically detect and download required ONNX models, fully offline local inference

      📝 Multiple Output Formats

      Supports SRT subtitles, JSON, plain text, and more

      ⚡ ONNX High-Performance Inference

      Supports multi-threaded inference, auto-detects CPU core count for intelligent adaptation

      🔧 Dual Mode Usage

      Interactive mode + CLI mode, flexibly adapting to different scenarios

      📦 Zero Config, Out-of-the-Box

      Unzip and run, no complex environment setup required

      ⏳ Loading...