AI Talking Head Video Production

★ 7.7 · content

talking-head-production is a Claude Code skill that produces talking-head videos with AI avatars, lipsync, and voiceover using the inference.sh CLI (belt). The recommended model is P-Video-Avatar — built-in TTS with 30 voices across 10 languages, generating at ~1.83 s/s for $0.025 per second of video, making it 18× faster and 6× cheaper than alternatives. The skill also covers OmniHuman 1.5 (multi-character conversations with up to two speakers), OmniHuman 1.0, Fabric 1.0, and PixVerse Lipsync. Workflows range from a single script-to-video command to full pipelines: portrait generation via P-Image, custom audio input, separate TTS via Dia-TTS, and multi-character dialogue. Ideal for spokesperson videos, online course content, social media clips, and product demos without any filming.