What you get

  • Photoreal hero stills and product scenes in assets/images
  • 3 to 5 second B-roll clips, including your approved still brought to motion
  • Original instrumental music beds and voiceover in 12 languages
  • One record per file in .media/manifest.jsonl, plus a provenance file with prompt, model, task and credits
  • A free dry run and one quoted total before anything is spent

Good for

  • Filling placeholder images and silent stretches in a HyperFrames film
  • Adding a music bed and voiceover to a launch, explainer or promo video
  • Voiceover cuts of the same film in several languages
  • Giving another video skill a media pass it can reuse

Not for

  • Interface, charts, type and logos: code or the owner's files do those better
  • Talking presenters, or a real person's voice or face without their consent
  • Projects that already have the media they need in assets/

How to use it

In Claude or ChatGPT

Use the PopNow skill "AI Assets for HyperFrames" (hyperframes-ai-assets) for this: My film in ./launch has a grey hero box and no sound. Add a hero still that moves a little, a music bed and an English voiceover. Quote it first.

In a coding agent

/popnow-hyperframes-ai-assets My film in ./launch has a grey hero box and no sound. Add a hero still that moves a little, a music bed and an English voiceover. Quote it first.

Then try

Add German and Japanese versions of the voiceover.

The bed is too busy. Make a calmer take and keep the old one.

What to have ready

  • A HyperFrames project folder
  • Node.js 22+ and Python 3
  • A PopNow API key (the dry run works without one)
  • The film's palette and mood, or its DESIGN.md

Similar skills

View all →

Questions

Does it work with the HyperFrames media skill?

Yes. Files land in the folders media-use reads, and each one is appended to .media/manifest.jsonl in the same record format, under the same lock, with the next free id, so reuse and candidate listing keep working. You can also have the HyperFrames CLI record them instead.

How do I know what it will cost?

A free dry run estimates every item, then a free quote gives the exact total and your balance. Generation starts only with that total confirmed, and a run stops rather than go over it.

What happens to files I already have?

Nothing. The script never overwrites: a new still next to your hero.png becomes hero-2.png. Items already generated with the same request are skipped, so running the plan again never pays twice.

Can it clone a voice or use someone's face?

Not by default. Plan items that reference a real person's voice or likeness, or ask to imitate one, are refused unless they carry a consent record from you, which is kept with the file.