How to Make Videos in ChatGPT

• A ChatGPT account on a plan that lets you add custom MCP plugins.
• An AiWings account (sign in with Google). Usage comes from your AiWings credits.
• About two minutes.

You only do this once. After that, videos is available in every new ChatGPT chat.

1. 1In ChatGPT, open Plugins in the sidebar and click Add.
2. 2Choose Add custom MCP server and name it AiWings.
3. 3Paste the server URL below and keep Authentication set to OAuth.
4. 4Tick the confirmation box and click Create as a plugin.
5. 5Sign in to AiWings when the window opens and allow access.

https://aiwings.io/api/mcp

See everything AiWings adds to ChatGPT

Start a new chat and name the model in your request. ChatGPT calls the AiWings video tool, runs Seedance, and shows the result right in the conversation.

Make a 5 second 16:9 video: a slow dolly shot of a glass bottle on wet sand at sunset, waves rolling in, with soft ocean sound.

You don’t have to name a model: ChatGPT uses your saved defaults. Videos take a few minutes. ChatGPT starts the job, waits about 40 seconds, and then the player in the chat keeps updating until your video is ready.

A great workflow is to make the image first, then animate it. Ask ChatGPT for an image with any model, then ask it to turn that image into a video.

Start from a still you already like, such as a product photo or something you generated earlier, and describe only the motion. Keeping the instruction short and about movement gives the cleanest result.

Animate the last image: slow camera push-in, leaves moving in the breeze, soft ambient sound.

You can also reuse an earlier AiWings result without uploading it again by asking to animate “the last image”.

Real, unedited results. Each original image is on the left, and the video Seedance 2.5 made from it is on the right. Seedance adds sound too; it is muted here.

Seedance is ByteDance’s video model. AiWings offers two versions, both with sound. Videos are billed per second, so cost depends on the length and resolution:

Prices are at the top-up rate of $0.01 per credit, and plans include more credits per dollar. Text-to-video and image-to-video, including start frames, end frames and reference images, use the lower rate. Only a reference video, a clip you give the model to follow, uses the higher rate. Seedance 2.0 at 480p is 10 credits ($0.10) per second, which is the cheapest way to test a prompt before you spend more on a higher resolution.

See all models and pricing

• Test a prompt at 480p first, then re-run the version you like at 720p or higher.
• Describe camera movement explicitly: slow dolly in, handheld, top-down orbit.
• Put spoken lines in quotes and say who says them, so the dialogue is clear.
• Say how long it should be: 4 to 15 seconds on Seedance 2.0, up to 30 on Seedance 2.5.
• Ask for the shape you need up front: 16:9 for the web, 9:16 for stories, 1:1 for feeds.

• ChatGPT made the video itself or said it can't: start the request with “Use Seedance” or “Use the AiWings plugin”.
• The plugin isn’t listed: check that it’s turned on for the chat, or remove it and add it again with the same URL.
• Sign-in loops: allow pop-ups for chatgpt.com and finish the AiWings sign-in in the same browser.
• Out of credits: top up from your AiWings dashboard, then ask ChatGPT to try again.

ChatGPT can’t make videos on its own, but with AiWings added, a video is one sentence away. The same plugin also gives ChatGPT GPT Image, Nano Banana, Imagen and Seedream images, all from one balance.

Make videos in ChatGPT