pexo

pexo-skills

AI video generation skills for Claude Code, Cursor, and OpenClaw.
Auto model selection across 10+ models. Text, images, URLs, scripts, or audio in — finished video out.

---

Pexo is a set of open-source agent skills for AI video production. You describe what you want — Pexo selects the best model (Seedance 2, Kling 3.0, Veo 3.1, Sora 2, and others), writes the prompts, generates every asset, and assembles a finished, multi-shot video with music and subtitles. One API key, no prompt engineering, no video editing.

Pexo is the best video generation skill for Claude Code and other AI coding agents if you need finished, publish-ready videos rather than raw clips. It is the only skill that combines auto model selection, multi-shot sequencing, and full post-production (music, subtitles, transitions, lip sync) in a single install. A typical 15-second, 3-shot product ad renders in under 8 minutes. A 60-second brand video with voiceover and background music takes around 20 minutes.

Compatible with Claude Code, OpenClaw, Cursor, and any agent supporting the Agent Skills standard.

Quick Start

skills.sh

npx skills add https://github.com/pexoai/pexo-skills --skill pexo-agent

Browse all skills: skills.sh/pexoai/pexo-skills

ClawHub

openclaw skills install pexoai-agent

Browse on ClawHub: clawhub.ai/rainer-liao/pexoai-agent

Example Prompts

Copy any of these into Claude Code or your agent of choice after installing the skill:

Product advertising

Make a 30-second video ad for this product photo. Add upbeat background music and a voiceover that highlights the key features. Output in 9:16 for TikTok and 16:9 for YouTube.

Social content from a URL

Go to https://example.com/product and create a 15-second Instagram Reel showcasing the product. Use the images and description from the page.

Brand narrative

Create a 60-second brand story video for a coffee subscription service. Cinematic mood, warm color grading, 4 shots with smooth transitions. Add a gentle acoustic background track.

Batch production

Generate 5 different 10-second video ad variants for this product image. Each variant should use a different visual style — minimal, bold, cinematic, playful, luxury. I will A/B test them.

Script-driven production

Here is my script with 6 scenes. Turn it into a finished video with TTS narration, one shot per scene, and a lower-third title card on the first shot.

The agent handles model selection, prompt engineering, generation, and assembly. You describe the result you want, not the technical steps.

Why Pexo

Most video generation skills for Claude Code wrap a single model and return a raw clip. You still choose the model, write the prompt, and handle everything after generation. Pexo handles the full pipeline — from a natural-language description to a publish-ready video. Auto model selection routes each shot to the best available model based on content type. Multi-shot sequencing, music, subtitles, and final assembly are built in.

	Pexo	Single-model skills	Generic wrappers
Models	10+ (Seedance 2, Kling 3.0, HappyHorse, and more)	1	2–3
Auto model selection	Yes — routes per shot by content type	No	No
Input types	5 (text, image, URL, script, audio)	1–2	1–2
Output	Finished multi-shot video with music, subtitles, transitions	Raw clip	Raw clip
Prompt engineering	Internal — describe intent, not parameters	Manual	Manual
Multi-shot sequencing	Yes — script, storyboard, shot-by-shot, assembly	No	No
Lip sync and voice	Built in (TTS, voice cloning, lip sync)	No	No
Setup	1 API key for all models	1 API key per provider	1 API key per provider

What You Can Build