Product Content & Copy AI

AI Video Generation

AI Video Generation is an ecommerce AI skill for 101 Skills, built for teams working with Codex, Claude Code. Use it to you need a repeatable way to produce product,…

Provider
101 Skills
Platforms
Codex · Claude Code
View original link · Skills.sh

What this skill helps you do

A CLI-driven skill for generating videos with 40+ AI models (Veo, Seedance, and more) through the inference.sh belt CLI. Covers login, model selection, prompt input, and output handling so you can produce product clips, ad creatives, and social videos directly from the terminal.

Install and get started

Copy the full instructions into your AI tool. Test one low-risk example before connecting real store data.

Original Skill instructions

You are an AI video generation operator working through the inference.sh belt CLI. Prerequisites: belt CLI installed and `belt login` completed. To generate video: 1) Clarify the goal (product clip, ad creative, social cut), aspect ratio, and duration; 2) Pick a model (e.g. google/veo-3-1-fast for speed, higher-quality models for hero shots); 3) Draft a cinematic prompt describing subject, motion, camera movement, lighting, and style; 4) Run `belt app run <model> --input '{"prompt": "..."}'`; 5) Review output, iterate on motion and camera language rather than re-rolling blindly. Keep prompts concrete: one subject, one camera move, one mood.

Useful tasks

  • Generate a 5-second product hero clip for a landing page
  • Produce short ad creatives for Meta/TikTok campaigns
  • Create lifestyle motion shots without a video shoot
  • Prototype video concepts before commissioning production

How to use it

  • Start with fast models for drafts, switch to quality models for finals
  • One camera move per prompt — compound moves confuse the model
  • Describe motion explicitly (dolly-in, orbit, static tripod)
  • Budget iterations: expect 3-5 runs per usable clip

More skills for this workflow

Content checked: