The Model Context Protocol is how assistants like ChatGPT and Claude use other apps. Add Sceneform once and your assistant can make images, videos, and speech with your brand, and schedule posts, without leaving the chat.

Does it use my credits?

Yes, the same credits as the app, and some things cost less: your assistant does the planning and writing itself, so you don't pay for Sceneform's AI to do it. A generation starts as soon as your assistant asks for it, after your assistant's own permission prompt, and shows what it cost. Tell your assistant to check with you first and you get a quote; nothing is charged until you approve it, and an approved quote is never charged twice.

Can my assistant post without asking?

No. A new post always waits for you to approve it. Before your assistant edits, cancels, or publishes a scheduled post, the app it runs in asks you first.

Where do the videos and images go?

Into your Sceneform library, the same as anything you make in the app. ChatGPT and Claude also show them in the chat.

Which assistants work?

ChatGPT, Claude, Claude Code, Codex, Cursor, Gemini CLI, and any other assistant that connects to remote MCP servers with sign-in. The sceneform CLI works from any terminal.

Remove Sceneform in your assistant's settings. In ChatGPT that is the Plugins page; in Claude, Customize, then Connectors.

Your next post,made in Claude Code.

Ask for a UGC video, an ad, or a product photo. Sceneform makes it with your brand, products, and creators, and posts it when you say so.

Paste this into Claude Code

Set up Sceneform for me so I can make videos, images, and posts from here.
1. Run `claude mcp add --transport http --scope user sceneform https://sceneform.ai/api/mcp`.
2. Run `npx skills add https://sceneform.ai -g -a claude-code -y`.
3. Tell me to run /mcp, pick sceneform, and sign in. Then call get_credit_balance and show me my plan and credits.

https://sceneform.ai/api/mcp

Ask for the job.Get the finished post.

A man in his kitchen cracks a tangle can that foams over, with captions and a 0g sugar sticker

A woman blends Solenne tinted moisturizer into her cheek, with SPF 30 and sheer tint callouts

A woman on a city sidewalk tries a tangle can in a blind taste test

Open in ChatGPT

A brand film in Motion Studio

@Sceneform A 30 second launch film for our brand. Close-ups, the faces from our posts, the line 'none of it was filmed', and our logo at the end.

Sceneform BrandKit

Open in ChatGPT

A skateboarder carving down a rain-soaked neon street at night

A man in a break room tied up in green ribbon sips a tangle can, the knots fall away and the room turns orange

A woman lifting the dropper from a green bottle of Lumé serum

Espresso being poured into a green cup in a warm café

One message.One finished video.

Not a generic AI video.Yours, from the first message.

Your brand

Give Sceneform your website once. Your logo, colors, fonts, and voice go into everything your assistant makes.

An orange tangle can on a plain backdrop

Your products

Your real products, from your store or a few photos. Say "our tangle can" and that can shows up.

Your creators

Save a face once and use it in every UGC video, so your posts have the same people week after week.

Your assistant asks.You decide what gets made and posted.

Jon, a saved creator, in his kitchen — Ask to check first and you get a quote. Nothing is charged until you press Generate.

A man in his kitchen holds up a tangle can, cracks it, and takes a sip — The video plays right in the chat and lands in your Sceneform library.

The generated video of Jon holding up the tangle can — A post always waits for your yes: the video, the account, the time, the caption.

FAQ

Not here? Email [email protected].

Your assistant already writes the caption.Now it makes the post too.