What problem does it solve? Producing monetizable short-form videos requires voiceovers, sound effects, and background music that are commercially license-safe, plus a properly mixed audio track — sourcing and mixing these manually is slow and legally risky. ## Core Features & Use Cases - Open-source TTS voiceovers: Generate narration with Kokoro-82M (Apache 2.0), Piper, or Chatterbox, including a deep-voice recipe using pitch shifting and warmth EQ on CPU. - License-safe audio sourcing: Pull SFX and music from CC0/CC-BY sources like Kenney, Pixabay, Freesound, and Incompetech with attribution rules for monetized content. - Deterministic FFmpeg mixing: Duck music under narration via sidechain compression, hit -14 LUFS loudness, and integrate audio tracks into HyperFrames compositions. - Audit gate: Run audit-audio.mjs for automated license and mix-contract checks, then a fresh audio-auditor subagent scores the mix (/50, PASS/FIX NEEDED) before delivery. - Use Case: Add a deep male Kokoro voiceover, a ducked Pixabay music bed, and CC0 whoosh SFX to a 30-second reel, then verify the mix passes the audit harness. ## Quick Start Use the voice-sfx-audio skill to generate a Kokoro voiceover for my reel script, add royalty-free background music, and mix everything with FFmpeg ducking at -14 LUFS.
