P Video Avatar

Generates talking head videos from a single portrait image using Pruna's P-Video-Avatar model through the inference.sh CLI. You give it a photo and either text (uses built-in TTS with 30 voices across 10 languages) or your own audio file, and it creates a lip-synced video at 720p or 1080p. Pricing is $0.025 per second of output for 720p, which is legitimately cheaper than HeyGen or Synthesia. Works well for product demos, multilingual content, or any situation where you need a consistent AI presenter without filming. You can control speaking style and background behavior through prompts. The workflow is straightforward: run the command with an image URL and script, get back a video. Pairs nicely with their P-Image model if you need to generate the portrait first.

skills.sh installs: 2.4K
GitHub stars: 437
last updated: May 16, 2026

View on GitHub

Install

$ npx -y skills add inference-sh-skills/skills --skill p-video-avatar --agent claude-code

Installs into .claude/skills of the current project.

Files

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

Pruna P-Video-Avatar

Generate talking head avatar videos from a single portrait image via inference.sh CLI.

P-Video-Avatar is the fastest and most cost-effective avatar video model available. Quality on par with Veo 3.0, 18x faster and 6x cheaper than alternatives like Fabric, OmniHuman, and HeyGen.

Quick Start

Requires inference.sh CLI (belt). Install instructions

belt login

# Generate avatar from portrait + text script
belt app run pruna/p-video-avatar --input '{
  "image": "https://portrait.jpg",
  "voice_script": "Hello, welcome to our product demo!",
  "voice": "Zephyr (Female)"
}'

Full Workflow: Generate Portrait + Avatar Video

Use Pruna P-Image to generate the portrait, then P-Video-Avatar to animate it:

# 1. Generate a portrait image with P-Image
belt app run pruna/p-image --input '{
  "prompt": "professional headshot portrait of a young woman, neutral background, looking at camera, studio lighting, photorealistic",
  "aspect_ratio": "9:16"
}'

SKILL.md on GitHub

Details

Repository: github.com/inference-sh-skills/skills
First seen: May 16, 2026