p-video-avatar — Claude Code skill: talking head avatar videos
This skill turns a portrait into a talking video through Pruna P-Video-Avatar and the inference.sh CLI. Handy for demos, explainers and localization.
npx skills add 101-skills/superpowers --skill p-video-avatar -a claude-codeWhat p-video-avatar does
The skill covers Pruna P-Video-Avatar, driven by the belt CLI from inference.sh. You feed it a portrait image and a line of text, and get back a video of that person speaking. Speech synthesis is built in, so no separate TTS service is involved.
There are 30 voices and 10 languages, at 720p or 1080p. You can pass your own audio file instead of text; when both are present, audio wins over voice_script. The video_prompt and voice_prompt parameters steer avatar behaviour and speaking style, and seed makes a run reproducible. If you have no portrait yet, P-Image can generate one.
You need an inference.sh account and a belt login before the first run. Pricing scales with length: $0.025 per second of output at 720p, $0.045 at 1080p. That adds up on long clips, so budget it before you queue a batch. The author suggests keeping videos under 3 minutes, otherwise the face drifts from frame to frame.
What it helps with
- turn a single photo into a talking avatar video
- voice a script without wiring up a separate TTS service
- dub the same clip into several languages
- generate a portrait and animate it in one pass
- produce a presenter clip for a product demo
Example prompts for the agent
- Make a talking avatar video from
portrait.jpgwith the line "Hi, this is our product demo." - Run the same portrait with
speech.mp3instead of a text script. - Generate a portrait with P-Image, then animate it at 1080p.
How to install p-video-avatar
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill p-video-avatar -a claude-code2Manually: put the files in place
~/.claude/skills/p-video-avatar/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/p-video-avatar/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill p-video-avatar -a codex2Manually: put the files in place
.agents/skills/p-video-avatar/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill p-video-avatar -a cursor2Manually: put the files in place
.cursor/skills/p-video-avatar/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install 101-skills/superpowers p-video-avatar2Manually: put the files in place
.github/skills/p-video-avatar/SKILL.md1Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill p-video-avatar -a gemini-cli2Platform CLI
gemini skills install https://github.com/101-skills/superpowers --path tools/video/p-video-avatarManually: put the files in place
.devin/skills/p-video-avatar/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add 101-skills/superpowers --skill p-video-avatar -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about p-video-avatar
How do I install p-video-avatar in Claude?
Run npx skills add 101-skills/superpowers --skill p-video-avatar -a claude-code. The skill lands in ~/.claude/skills/p-video-avatar/SKILL.md.
What does p-video-avatar do?
It generates talking head avatar videos from a portrait image via Pruna P-Video-Avatar and the inference.sh CLI. Speech comes from the built-in TTS, or you can supply your own audio file.
What do I need before using it?
An inference.sh account and a belt login. Billing runs per second of output video: $0.025 at 720p and $0.045 at 1080p.
