Connect Frame24 once
In ChatGPT's settings, turn on Developer mode, create an app named Frame24 and sign in. About three minutes, no code; Plus, Pro, Business, Enterprise or Education on chatgpt.com.
ChatGPT video generator
Can ChatGPT make videos?
Not since Sora closed.
ChatGPT's video generation was Sora, and OpenAI shut it down in 2026: the app in April, the API in September. ChatGPT still writes the script and the shot list. To get a finished product video from inside the chat, connect a tool that makes one.
Three 30-second films · Card required · No generated media
In plain terms
- What Frame24 is
- ChatGPT no longer generates video. Frame24 is an AI product video generator that runs inside ChatGPT, Claude and coding agents as an app: you ask in the conversation and get back a narrated film of your product, made from your website, brand and real screens.
- Best for
- Founders, marketers and developers who searched for a ChatGPT video generator because they need a video of their product, and want to keep working in the assistant they already use.
- How it differs
- A general text-to-video model makes clips from a prompt and knows nothing about your product. Frame24 reads your website, captures your running app and writes the story from what the product actually does, so the film can only be about you.
What happened to Sora
Sora's shutdown, date by date.
OpenAI closed its video generator in two stages. Since then no part of ChatGPT renders a video.
| Date | What changed |
|---|---|
| 24 March 2026 | OpenAI announces that Sora will shut down |
| 26 April 2026 | The Sora app and website close; ChatGPT stops generating video |
| 24 September 2026 | The Sora 2 models and the Videos API are shut down, with no one-to-one replacement |
Sources: OpenAI video generation guide and The Decoder, checked 1 November 2026.
What ChatGPT still does
ChatGPT writes the video. Something else makes it.
The language half of a video is still ChatGPT's strength. Use it for that, and hand the result to a tool that renders.
| Job | ChatGPT today | What makes the video |
|---|---|---|
| Script and voiceover text | Yes | Your video tool's narrator, or a recording |
| Storyboard and shot list | Yes | The scenes your video tool builds |
| Still images and thumbnails | Yes | Used as frames or covers |
| A rendered video file | No | A video tool, connected as an app or used on its own |
| A video of your own product | No | A product video tool that can see your website and real interface |
Make the video from ChatGPT
A product film, without leaving the chat.
Frame24 is a ChatGPT app. Once it is connected, you ask for the film in plain words and ChatGPT checks each step with you before anything is made.
Ask, and show the product
Say what you need: a 30-second homepage film, a launch video, a sales follow-up. Frame24 reads your website and brand; sign in to your app once in its secure browser and the screens that prove the story are captured.
Approve the story, get the film
Review the storyline before it is made, then get a narrated film with music and captions, and change any scene by describing it.
Each step with pictures: connect Frame24 to ChatGPT. It also works in Claude and Gemini.
Claude Code, Codex and Cursor
Claude Code video, Codex video: the agent makes the film.
A coding agent already knows your product, because it can read the code and run the app. That is the hard part of a product video.
| Route | What you get | Who directs |
|---|---|---|
| Write the animation as code (Remotion) | Total control over every frame | You: story, animation, sound and timing |
| Drive a video service over MCP (Frame24) | A finished, narrated film from your real interface | You approve the story; the service makes the film |
| Prepare a recording | Script and screens for a screen recording you make | You, on camera or on screen |
Commands for each agent: Frame24 for coding agents. The three routes in depth: can coding agents make product videos?
Finished films
Made from a website and a product.
Each began with a company's own website and real interface.
Use Frame24 from your AI assistant
Let the agent that already knows your work make the film.
Connect Frame24 to ChatGPT, Claude, Gemini or a coding agent in about two minutes. Your assistant can understand the product, capture the right screens, review the film and keep the whole process in the conversation.
Choose your assistant and connect itclaude mcp add --transport http frame24 https://studio.frame24.app/api/mcpStart Claude Code, type /mcp, choose frame24 and Authenticate. Sign in to Frame24 in the browser tab that opens and click Allow.
Questions
Answered plainly.
Can ChatGPT generate videos?
Not any more. ChatGPT's video generation came from OpenAI's Sora, and OpenAI closed the Sora app and website on 26 April 2026 and shut down the Sora 2 models and the Videos API on 24 September 2026, with no one-to-one replacement. ChatGPT can still write a script, a storyboard and shot prompts, and make still images, but it does not render a video file itself.
What happened to Sora?
OpenAI announced the shutdown on 24 March 2026, closed the consumer app and website on 26 April 2026, and turned off the API on 24 September 2026. It did not give a detailed public reason.
How do I make a video with ChatGPT now?
Use ChatGPT for the words and connect a tool that makes the video. For a product or software video, Frame24 runs inside ChatGPT as an app: you ask in the chat, it researches your product, captures your real screens and returns a finished, narrated film. For cinematic or generic clips, use a dedicated text-to-video tool and bring ChatGPT's script and prompts to it.
Which ChatGPT plans can use the Frame24 app?
ChatGPT Plus, Pro, Business, Enterprise and Education, on chatgpt.com in a browser. Custom apps are not available on the free plan or in ChatGPT's mobile apps yet. Setup takes about three minutes and needs no code.
Can Claude Code or Codex make a video?
Yes, two ways. A coding agent can write the animation as code with Remotion and render it, which gives you full control and makes you the director. Or it can drive a video service over MCP: connected to Frame24, Claude Code, Codex or Cursor reads your product, captures the real interface from your running app and gets a finished film back.
What does Frame24 cost?
Frame24 starts with a 3-day trial (card required) that includes three 30-second films. Downloading one finished 30-second film is $20, and plans start at $39 a month. There is no free plan.
Ask for the film
in the chat.
Connect Frame24 to ChatGPT, Claude or your coding agent and ask for a product film. Three 30-second films on a 3-day trial.
Start your trial Three 30-second films · Card required · No generated media