Turn any image into an 8-second video from inside Claude.

A Claude Code skill that reads your image, drafts a cinematic prompt, lets you edit it conversationally, then ships an MP4 via Google Veo 3.1. Built for marketers building paid-ad creative.

What it does

• -Reads the image you share. Drag-drop, URL, or local path. No prompt engineering required.
• -Drafts the Veo prompt for you. Claude looks at the image and generates a cinematic prompt with motion, framing, and lighting.
• -Iterates conversationally. "Make it slower," "add more pull-back" — refine the prompt before spending generation credits.
• -Outputs an 8-second 9:16 MP4. Saves to your Downloads folder, ready to drop into a paid-ad placement.

Who it's for

Marketers, creators, and brand teams running Claude Code who want iterative creative control before spending Veo credits. If you're running paid ads on Meta, TikTok, or YouTube and need fast 9:16 video iterations from product photos, this is for you.

How it works

1. 01Install once. Clone the skill into your Claude skills folder. Restart Claude.
2. 02Add a Gemini API key. Free key from Google AI Studio. Veo 3.1 must be enabled on the project — some accounts need to request access.
3. 03Send Claude an image. Paste a Dropbox or public URL, drop a local path, or attach the file directly to the message.
4. 04Approve or edit the prompt. Claude shows the proposed Veo prompt with motion, framing, and lighting. Tweak it. When you're happy, ship it.
5. 05Get the MP4. ~30s wait. File lands in ~/Downloads with a sensible name. Re-runs auto-suffix so you never overwrite a previous take.

Get the install

Drop your email and the GitHub link plus the exact install commands appear below. The skill itself is free, the repo is public, and one email unlocks every gated thing on this site.