skills.rest
genmedia-labs
genmedia-labsOFFICIAL
Explore all skills by this author →
GitHub

ai-avatar-video

Generate talking-head and lip-sync avatar videos via the RunComfy CLI.

12|2|Updated Aug 12, 2026
Search Tags:#text-to-video#talking head#digital human#lip sync#omnihuman#avatar video#runcomfy
One-click install
npx skills add https://github.com/genmedia-labs/skills --skill ai-avatar-video-genmedia-labs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-avatar-video
Source: https://github.com/genmedia-labs/skills/tree/main/ai-avatar-video
Command: npx skills add https://github.com/genmedia-labs/skills --skill ai-avatar-video-genmedia-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @runcomfy/cli.

What problem does it solve? Creating AI avatar, talking-head, and lip-sync videos requires choosing between many audio-driven video models with different input requirements. This Skill routes each request to the right RunComfy model and provides the exact CLI invocation for it. ## Core Features & Use Cases - Intent-based model routing: Selects among OmniHuman, Wan 2-7 with audio_url, Wan 2-2 Animate, HappyHorse 1.0, and Seedance v2 Pro based on whether the user has an audio file, a script, a photoreal portrait, or a stylized character. - Ready-to-run CLI commands: Ships documented prompting patterns and minimal runcomfy run invocations for each model route. - Use Case: A marketer needs a UGC-style product ad with a vertical talking-head video from a portrait photo and a voiceover MP3. The Skill routes to OmniHuman and produces the video with a single runcomfy run call. ## Quick Start Ask the agent to make the person in your portrait image speak the lines from your audio file using the ai-avatar-video skill.

Frequently Asked Questions about ai-avatar-video

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make a talking head video from a photo and audio file?▼

Use the OmniHuman route: run `runcomfy run bytedance/omnihuman/api` with an `image_url` for the portrait and an `audio_url` for the voiceover. The model generates a video where the subject speaks or gestures naturally to the audio, with no prompt required.

Which AI avatar model should I use for lip-sync video?▼

Use OmniHuman for a portrait plus pre-recorded audio, Wan 2-7 with audio_url for full scene control with locked lip-sync, HappyHorse 1.0 when you only have a written script, Wan 2-2 Animate for stylized characters, and Seedance v2 Pro for cinematic multi-reference shots.

Can I create a talking avatar video without an audio file?▼

Yes. HappyHorse 1.0 generates speech in-pass from a written script. Quote the spoken line inside the prompt using `says clearly: "..."` and optionally describe the audio tone separately, such as a friendly tone with a neutral accent.

Does OmniHuman work with stylized or anime characters?▼

OmniHuman is designed for photoreal portraits. For illustrated, anime, or mascot characters with full-body audio-driven motion, use the community Wan 2-2 Animate model instead, which accepts a character image and an audio file.

Why is my lip-sync video mouth movement inaccurate?▼

Audio quality drives output quality. Clean voiceover without a music bed produces cleaner mouth sync, so isolate the voice stem first if your audio is a mix. Also avoid putting spoken words in the prompt for audio-driven models, since they sync to the waveform.

Related Skills

Explore More Design & Creative Skills→
prime-skillsprime-skills
31

ai-avatar-video

Create talking-avatar and lip-synced videos from portraits, scripts, and audio files.

Official
Advanced
VictaaawVictaaaw

video

Generate marketing videos with AI avatars, voiceovers, and programmatic templates.

Community
Advanced
prime-skillsprime-skills
31

face-swap

Replaces faces or characters in images and videos via RunComfy APIs.

Official
Advanced
Pixazo-AIPixazo-AI

veed

Generate talking-avatar, lip-synced, and background-removed videos via the Veed API on Pixazo.

Official
Intermediate