+
CHATGPT + AI BOX
Generate videos inside ChatGPT
ChatGPT can't generate videos natively — OpenAI discontinued Sora in 2026. The AI Box connector brings video back. Add one MCP connector in ChatGPT's settings and ChatGPT can generate videos with Google's Veo 3, ByteDance's Seedance, and PixVerse — text-to-video and image-to-video, right inside your conversation. No API keys, no code.
HOW IT WORKS
Live in three steps
Turn on Developer mode
In ChatGPT on the web: Settings → Apps & Connectors → Advanced settings → Developer mode (paid plans).
Create the AI Box connector
Click Create, paste the AI Box MCP URL, keep OAuth authentication, and approve the popup.
Enable it in a chat and ask
Click + in the composer, toggle AI Box on, and describe what you want to create.
THE MODELS
ChatGPT picks the tool. You pick the result.
Google's Veo 3, ByteDance's Seedance, and PixVerse V5 — every top video model on AI Box becomes available to ChatGPT the moment you connect.
Veo 3
Seedance
PixVerse+ more added monthlyTRY IT
Prompts that just work
PRODUCT
"Use AI Box to generate a 5-second video of a coffee cup steaming on a windowsill at sunrise, cinematic lighting."
SOCIAL CLIP
"Make a vertical 9:16 clip of ocean waves in slow motion I can use as a background for a quote post."
IMAGE-TO-VIDEO
"Take the product image I just generated and animate it into a slow camera push-in for our landing page hero."
WHAT PEOPLE MAKE
- Short clips for social while ChatGPT writes the caption
- Landing page hero and background loops
- Animating a generated image into motion
- Ad creative variations to test
- B-roll for videos and presentations
FAQ
Questions, answered
Can ChatGPT generate videos?+
Not natively — OpenAI discontinued Sora in 2026, so ChatGPT has no built-in video generation. But with the AI Box MCP connector, ChatGPT can generate videos using models like Google's Veo 3, ByteDance's Seedance, and PixVerse, directly inside your chat.
What happened to Sora?+
OpenAI announced Sora's discontinuation in March 2026. The consumer app closed in April 2026 and the API is scheduled to shut down in September 2026. If your workflow depended on Sora, the same text-to-video and image-to-video jobs run through AI Box with Veo 3, Seedance, and PixVerse — inside ChatGPT, Claude, or the AI Box Playground.
How do I add video generation to ChatGPT?+
In ChatGPT on the web, enable Developer mode under Settings → Apps & Connectors → Advanced settings, then click Create and paste the AI Box MCP URL: https://mcp.aibox.ai/mcp/. Authorize with one click, enable AI Box in your chat's composer, and ask for the video you want. Custom connectors require a paid ChatGPT plan.
Which video models can ChatGPT use through AI Box?+
All the video models on AI Box — including Veo 3 and Veo 3 Fast, Seedance 1.0 Lite and Pro, and PixVerse V5. Describe the clip in plain English and the right model generates it, or ask for a specific model by name.
Can ChatGPT turn an image into a video?+
Yes. Through AI Box, ChatGPT can run image-to-video workflows — generate or upload an image, then animate it into a clip. You can chain the steps in one conversation: create the image, review it, then ask ChatGPT to bring it to life.
Where do my generated videos go?+
Videos link directly in your ChatGPT conversation, and every asset is also saved to your AI Box media storage — so a long render is never lost if a chat is deleted or a tool crashes mid-run.