ChatGPT is excellent at writing scripts, hooks and shot lists, and it can make images. Turning those into finished video inside the chat is a different job: it needs a video model. The simplest way to get that in ChatGPT is an app (built on MCP) that runs video models and returns the clip to the conversation.
Your options
- Write in ChatGPT, render elsewhere: copy prompts into a separate video site. Works, but you juggle tabs and files.
- Use an app inside ChatGPT: ChatGPT calls the video model for you, shows the price and the result, and keeps editing in the same thread.
Adding a video app to ChatGPT
- Turn on developer mode in ChatGPT settings (Apps & Connectors), where available on your plan.
- Create an app and paste your Typeshoot connector URL; choose OAuth and sign in.
- Start a new chat and ask for a clip.
A good first prompt
PromptWrite three hooks for a 15-second vertical ad for my cold brew, pick the best, then make it as a video with sound. Show me the cost first and start with a cheap draft.
From there you can ask to reframe it, add captions, or join several clips with music: the same editing tools work in Claude and ChatGPT.
Which model will it use?
Typeshoot includes a model advisor that reads your brief (length, shape, sound, speech, budget) and recommends from 16 video models, so you don’t need to know model names. For cinematic work it often picks Seedance; for fast social clips, Gemini Omni Flash. More on Seedance.
Make videos in ChatGPT
Typeshoot adds image, video, voice and editing tools to Claude, ChatGPT and other assistants. 100 free credits to start, no card needed.
Get 100 free creditsFrequently asked questions
Can ChatGPT generate videos?
ChatGPT itself focuses on text and images. To make videos inside a chat, connect an app (an MCP server) that runs video models and returns the clip.
Which video model should I use?
It depends on the brief: a fast all-rounder with sound for social clips, a cinematic model for ads, a cheap model for drafts. A model advisor can pick by length, shape, sound and budget.