ThePlasmak

Local speech-to-text using faster-whisper, offering 4-6x faster transcription than OpenAI Whisper with identical accuracy, and ~20x realtime with GPU acceleration. It includes subtitle generation, speaker diarization, URL/YouTube input, batch processing, and transcript search.

Speech-to-text with word-level timestamps, speaker diarization, and forced alignment using WhisperX. Built on faster-whisper with batched inference for 70x realtime speed.