How to make a product video from Claude or ChatGPT

What you need

• A LastCut account with credits in its workspace. Credits are bought on lastcut.video, never in the chat.
• Claude (web or desktop, any plan) or ChatGPT with developer mode turned on (Plus, Pro, Business, Enterprise or Edu, on the web).
• A product to make a video of: its product page link, links to its images, or photos on your device.

1. Connect LastCut to your assistant

LastCut is an MCP server, so it is added the way any custom connector or app is. In Claude, open Customize, then Connectors, press the plus, choose Add custom connector, name it LastCut and paste the server URL. In ChatGPT, turn on developer mode under Settings, open Plugins, create an app named LastCut with the server URL and OAuth authentication, leaving the client fields empty. Either way you then sign in to LastCut, choose the workspace whose credits the assistant may use, and press Allow.

The server URL and the menu-by-menu steps for both hosts — with links to Anthropic's and OpenAI's own help pages for when a menu moves — are on the connector page.

2. Switch it on for the chat

In a new chat, open the plus menu and switch LastCut on (Claude: Connectors; ChatGPT: Developer mode). Naming it in your message helps the assistant reach for the right tool: "Use LastCut to…".

3. Hand it the product

Paste the product page link, or links to the product images. The assistant imports the product and its photos and tells you which photos it could use — photos that do not show a garment are skipped, and it says which.

Use LastCut to import https://yourstore.com/products/linen-midi-dress.

For photos on your phone or computer, ask the assistant to add product photos. Assistants cannot pass on files attached to a chat, so a card appears to drop the photos into, with a link that works for 30 minutes if you would rather upload from a browser.

Add product photos to LastCut — I'll upload them myself.

4. Ask for templates that fit

Say what the garment is and what kind of video you want. The assistant searches the library by gender, department, garment and kind of video — OOTD, street walk, GRWM, studio, close-up — and shows a picker of live templates with each one's length and price.

Show me LastCut OOTD templates for a women's linen midi dress, shot outdoors.

Pick one from the picker, or tell the assistant which. Ask it to describe a template if you want to know what happens in it before you choose.

5. Get the quote

Ask for a quote. The assistant prices one video: the template, the model (the template's own or one from the library), which of your photos replaces which piece in the reel, and the credits against your balance. Nothing is spent at this step; a quote holds its price for 30 minutes.

Quote that template with my dress, on a different model with darker hair.

6. Say yes

The video is made only when you approve the quote — say yes, or press Generate on the quote card. Approving the same quote twice makes one video and charges once. If the workspace is short of credits the assistant says so and points at the pricing page; once there are credits, say yes again.

Yes, make it.

7. Collect the video

A video takes about ten minutes. Ask the assistant whether it is ready and it hands you the finished video as a card you can play, with a link to its page on lastcut.video. If a video fails, its credits come back to the workspace.

Is my LastCut video ready?

Things worth knowing

• The assistant never sees your conversation — LastCut receives only what the tools are sent: a link, a search, a quote id.
• You can disconnect at any time from Connected apps in your LastCut account, or from the assistant's own connector settings.
• The same template with a different product is a different video; the same model across a range is a matching feed. Ask for both.
• Only use products and photos you have the rights to.