Remove references to removed features (image gen, LLM chat, TTS). Accurately describe current project state as a local whisper.cpp speech transcription tool with GPU acceleration.
Remove image generation, text-to-speech, and chat features. Keep speech transcription with Whisper.cpp, model manager, and settings. Support multiple audio formats (WAV, MP3, M4A, OGG, FLAC, WebM, AAC) via ffmpeg transcoding. Single-call API: upload → transcribe → download as txt/json/md.
truetype