Dictata: Local Whisper voice dictation for Windows, no data leaves machine

Dictata addresses a growing demand for local, privacy-preserving dictation tools. The application combines Whisper (OpenAI's open-source speech recognition model, run via whisper.cpp) with optional local LLM post-processing, eliminating the privacy trade-off inherent in cloud-based dictation services. The developer designed the tool to be self-contained—the Windows executable carries its own icon and version metadata, requires no DLL files, and uses the Vulkan loader supplied by the graphics driver, simplifying deployment. The feature set reflects practical workflow needs: hotkey activation (toggle or push-to-talk), automatic paste into the active application, continuous streaming mode, and silence-skipping via voice-activity detection all reduce friction. The built-in ggml model library with hardware-aware recommendations and HuggingFace integration lower the barrier to model selection. The MIT-with-Commons-Clause licensing permits free personal and self-hosted use while blocking commercial resale, a choice that protects the creator's work while remaining permissive for non-business users.