talking-head-production

Installation
Summary

AI avatar talking head videos with lipsync, TTS, and multi-character support via inference.sh CLI.

  • Supports OmniHuman 1.5 for multi-character conversations and gestures, OmniHuman 1.0 for single characters, and PixVerse for quick lipsync on existing video
  • Requires high-quality source portraits (min 512x512, ideally 1024x1024+) with frontal gaze, neutral expression, and head-and-shoulders framing for accurate animation
  • Generates dialogue audio via Dia TTS with speaker tags and emotion control, then pairs with portrait to create videos up to 30 seconds per clip
  • Handles long-form content by splitting into segments, generating individual talking head clips, and stitching them together; supports caption integration and multi-character workflows
SKILL.md

Install the belt CLI skill: npx skills add belt-sh/cli

Talking Head Production

Create talking head videos with AI avatars and lipsync via inference.sh CLI.

Quick Start

Requires inference.sh CLI (belt). Install instructions

belt login
Related skills

More from inferen-sh/skills

Installs
GitHub Stars
500
First Seen