How to turn a ChatGPT image into a video: animate GPT images step by step
Opens image-to-video with VEO 3.1 Lite selected. Upload the image, describe the motion and check the cost (from 3 credits) before you generate.
Animate your ChatGPT imageA lot of people now make the picture in ChatGPT and then look for a way to make it move. This guide covers the parts specific to ChatGPT images: getting the right file and shape out of ChatGPT, writing a motion prompt for the styles people usually make there (illustrations, 3D characters, toy-box figures, product mockups, portraits), and avoiding the common failures. For image-to-video in general, see the complete image-to-video guide.
Can ChatGPT turn an image into a video by itself?
Not with its image tool: it returns still images. For a real video clip with camera movement, physics and sound you need a video model. The usual workflow is therefore two tools: ChatGPT for the picture, an image-to-video generator for the motion. Your ChatGPT image becomes the first frame, and the video model predicts every frame after it.
Step 1: Get the original file out of ChatGPT
- Download, don't screenshot. Open the image in ChatGPT and use the download button. A screenshot adds borders, lowers resolution and often includes interface elements the video model will try to animate.
- Keep the file as it is. ChatGPT downloads are often PNG, which is fine. In our measurements of 360 real photo uploads, PNG inputs finished more often than JPEGs (why image-to-video fails).
- Check the size. The Viralance uploader accepts images up to 10 MB, which covers normal ChatGPT downloads.
- Don't upscale it with a filter app first. Over-sharpened edges and smoothed skin tend to flicker once they move.
Step 2: Get the shape right before you animate
In image-to-video mode the clip takes the shape of your picture; there's no separate aspect-ratio setting. ChatGPT typically makes square, portrait or landscape images, so plan for where the video will go:
| Where it will be posted | Ask ChatGPT for | If you already have a different shape |
|---|---|---|
| TikTok, Reels, Shorts | "vertical 9:16" or "tall portrait" | Crop to 9:16, or extend the sides with Expand |
| Instagram feed | "4:5 portrait" or square | Crop |
| YouTube, website header | "wide 16:9 landscape" | Extend the top and bottom with Expand |
ChatGPT may still return a 2:3 portrait instead of an exact 9:16. That's close enough for most uses; for an exact fit, crop in any photo editor or use Expand on Edit Image (3 credits), which paints in the missing edges (outpaint guide). Leave a little empty space around the subject either way: models need room to move a head, an arm or the camera.
Step 3: Upload it and choose a model
- Open Create Video and choose Image to Video.
- Drop the ChatGPT image into the upload box (or pick it from My Photos if you made it in Viralance).
- Choose a model and length, write the motion prompt (Step 4) and check the credit cost shown before you press Generate.
Which model to start with, from the live Viralance catalog:
| Model | Lengths | Credits at 720p | Good for |
|---|---|---|---|
| VEO 3.1 Lite | 4 / 6 / 8 s | 3 / 4 / 5 | Cheap first test, can add sound |
| Hailuo 2.3 | 6 s | 3 | Smooth camera moves and physics, no sound |
| Kling 3.0 Standard | 5 / 10 / 15 s | 8 / 14 / 22 | Characters and longer shots |
| VEO 3.1 Fast | 4 / 6 / 8 s | 6 / 8 / 12 | Cleaner detail, sound, dialogue |
Start with VEO 3.1 Lite or Hailuo 2.3 to see whether the picture animates well. If it does, re-run the same prompt on a stronger model only when you need more detail or a longer shot. All models are compared on the models page.
Step 4: Write a motion prompt, not an image prompt
The most common mistake is pasting the original ChatGPT image prompt into the video tool. The model can already see the picture. What it needs is: what moves, how the camera moves, and what must stay the same.
A simple structure: subject action + camera move + atmosphere + what stays fixed.
Short prompts fail more often. In our data, image-to-video prompts of 10 words or fewer finished 50% of the time, against 86% at 11 words or more. One or two full sentences is the sweet spot.
Prompts for popular ChatGPT image styles
Hand-drawn anime-style illustration:
Gentle breeze moves her hair and the grass, clouds drift slowly across the sky, she turns her head toward the camera and smiles softly. Slow push-in. Keep the hand-drawn illustration style and colors exactly as they are.
3D animated character (cartoon or family-film look):
The character blinks, looks up at the camera and waves with one hand, a small happy bounce. Static camera, soft studio lighting. Keep the 3D cartoon style, proportions and outfit unchanged.
Action-figure or toy-box image:
The camera slowly orbits the toy box on the shelf, a soft light sweep moves across the plastic window, the figure inside stays still. Keep all text and packaging exactly as shown.
Product mockup:
Slow push-in toward the product on the marble counter, soft morning light moves across it, a little steam rises from the cup beside it. Keep the product's shape, color and label exactly as shown. No text.
Realistic portrait you made of a fictional person:
She laughs, looks away and then back at the camera, a light wind moves her hair, slow handheld camera at eye level. Natural skin, keep her face and outfit unchanged.
Landscape or poster-style scene:
Clouds roll slowly over the mountains, the lake surface ripples, birds fly across the distance. Slow pan to the right. Keep the composition and colors unchanged.
If the model ignores the camera direction, use the camera control on the create page; it appends a plain camera sentence to your prompt.
Step 5: Review, then fix one thing at a time
Watch the clip at full size, then:
- Face or character changes? Shorten the clip, reduce the action ("small head turn" instead of "turns around") or try Hailuo 2.3 or Kling 3.0 Standard.
- Style turns photoreal? Say the style in the prompt and add "keep the illustration style."
- Text melts? Generated video rarely keeps lettering intact. Remove text from the image before animating and add it back in an editor (keeping labels readable).
- Nothing moves? Your prompt only described the scene. Add a clear action verb and a camera move.
Change one thing per retry, so you know what fixed it. Failed generations are refunded automatically; a result you don't like is not.
Why a ChatGPT image gets refused
Most image-to-video failures are content-policy refusals triggered by what is in the picture, not technical errors. With ChatGPT images, the usual causes are:
- It looks like a real, identifiable person, such as a celebrity look-alike or a photo-real image of a friend. Video models screen for impersonation and public figures.
- Anyone who looks under 18 in a context that the model's filters treat as risky.
- Brand characters and logos from films, games or sports.
Fictional people, your own products and original characters go through far more reliably. More detail and fixes are in why AI video prompts get rejected.
Other ways to bring a ChatGPT image to life
- One-photo templates. Upload the image to a template and get a fixed effect, such as a product spin, an unboxing or a seasonal scene, without writing a prompt.
- Make it talk. For a character speaking a script, use AI Actors or Lip Sync on Edit Video (AI spokesperson guide).
- Copy a real movement. To make a ChatGPT character dance or gesture like a reference video, use Motion Transfer.
- Make the image here instead. Create Image includes OpenAI's GPT Image alongside Nano Banana and Seedream (1–4 credits per image), and any result in My Photos has an Animate to video action.
Can I use a ChatGPT image commercially in a video?
OpenAI's terms of use say that, as between you and OpenAI, you own the output, as long as you follow its policies. That does not cover third-party rights: a picture that copies a trademarked character, a logo or a real person's likeness can still create problems. Viralance's terms say you keep ownership of what you generate. Platforms also ask you to label realistic AI content; TikTok, for example, requires a label on realistic AI-generated videos (TikTok Support).
Frequently asked questions
How do I turn a ChatGPT image into a video? Download the image from ChatGPT, upload it to an image-to-video tool, and describe the motion and camera move you want. On Viralance: Create Video, Image to Video, upload, choose a model, write the motion prompt and generate.
Can ChatGPT make videos from images? ChatGPT's image generator makes still images. To get a real video clip from one, use a separate image-to-video model with the picture as the first frame.
Why does my animated ChatGPT image look different from the original? The model invents every frame after the first, so faces, hands and text can drift. Keep clips short, ask for small movements and say what must stay unchanged.
What size should a ChatGPT image be for TikTok or Reels? Vertical 9:16. Ask ChatGPT for a tall vertical image, then crop or extend it to 9:16 before animating, since the video keeps the image's shape.
How much does it cost to animate one image? On Viralance, from 3 credits for a 4-second VEO 3.1 Lite or 6-second Hailuo 2.3 clip at 720p. New verified accounts get 4 free credits, and packs start at $9 for 30 credits with no subscription (pricing).
Can I add sound or a voice? VEO 3.1 Lite, VEO 3.1 Fast and Kling 3.0 can generate sound with the clip. For narration, add a voiceover from a script.
