voiceCLI

Pick a clean speech segment from any audio or video file for voice cloning — silence-based VAD, scoring by duration / position / edge-cleanliness, preview MP3s. Replaces blind `--start`/`--duration` guessing. Triggers: "sample pick" | "pick sample" | "clean segment" | "find speech" | "pick voice segment" | "extract clean audio".

VoiceCLI assistant — author TTS scripts, generate speech, clone voices, transcribe audio, manage samples. Knows engine capabilities, markdown format, and all CLI commands. Triggers: "voice" | "voicecli" | "speech" | "generate speech" | "clone voice" | "transcribe" | "TTS" | "text to speech" | "voice script".

Design and optimize a TTS voice personality through evolutionary optimization. Generates batches of profile variants, waits for ranking feedback, iterates until convergence, then writes the winning profile to voicecli.toml. Triggers: "design a voice" | "voice design" | "tune voice personality" | "voice personality" | "/voice-design".

Clone a voice from a YouTube video — download audio, auto-pick a clean segment (via sample-pick), import it into voicecli, manage VRAM, and generate a test clone. Triggers: "clone from youtube" | "voice from youtube" | "yt clone" | "clone youtube voice" | "extract voice from video".