Images, video and the media library
Replace any image or video, upload your own, generate new ones with AI, and manage everything you have generated in the Media panel.
For the complete documentation index, see llms.txt. This page is also available as Markdown.
#Replace an image or video
Step 1: Hover the media
On any image, background-image container or video, a Replace image or Replace video pill appears.
Step 2: Click the pill, or select the element and open the Images tab
The pill opens a small picker with Recent, Upload and Generate tabs. With an image selected, the Images tab shows Current Image, Upload, Paste URL and Generate with AI. The Replace button in the floating toolbar opens the Images tab.
Step 3: Choose the new file
Upload a file, paste an address, pick one from Other images in project, reuse a recent generation, or generate one.
#Upload limits
| Item | Limit |
|---|---|
| Images | Up to 10 MB each |
| Videos | Up to 50 MB each |
| Accepted types | Images, mp4, webm, quicktime, m4v, mov, heic and heif |
| Where videos can go | The pill's picker and the Content tab's Upload Media zone accept video. The Images tab's Upload from device accepts images only |
Uploads are stored at permanent addresses, not inline data. HEIC and HEIF photos are converted to JPEG in the editor, first in your browser and then on the server. If that fails you see This HEIC photo could not be converted — please export it as JPG/PNG and try again. The homepage generator rejects HEIC outright, so use JPG there.
The size errors read: Image is larger than 10 MB — please pick a smaller file., Video is larger than 50 MB — please pick a smaller / compressed file. and, in the Content tab, Video <name> is <size>MB — max 50MB. Compress it first. If an upload fails, a message reads Upload failed — your change was NOT applied. and the image stays as it was.
#The Images tab
| Control | What it does |
|---|---|
| Current Image | A preview, or "No image set" |
| Alt text | The description read by screen readers and shown if the image fails to load. The box reads "Describe the image…" |
| Object fit | Default, Cover, Contain, Fill or None. Applies at every size |
| Upload from device | Choose an image from your device |
| Paste URL | Use an image from a web address. Press Enter or Apply URL |
| Generate with AI | Choose a model, an aspect ratio, optionally a reference image, write a prompt and press Generate image |
| Recent generations | Up to 12 images you generated in this editing session. Not kept after you close the editor |
| Other images in project | Every other image on every page. Click one to use it |
With a non-image element selected, the tab shows Images in your project, a tip reading "Select an image on the canvas to upload, paste a URL, or generate with AI.", and thumbnails that select the matching image when clicked.
#Generate with AI
| Setting | Options |
|---|---|
| Model | Imagen 4 (Best quality, recommended, the default), Seedream 5 (Premium, editorial), FLUX 2 Pro (Reliable, versatile), FLUX Schnell (Fastest, drafts) |
| Aspect ratio | 16:9 (the default), 4:3, 1:1, 3:4 |
| Reference image | Optional. Add reference attaches one image to guide the result |
Each generation is charged in credits at the model's price, and a failed request shows "Generation failed" with its status code. See What actions cost and Model pricing.
#Background images and video
A container with a CSS background image shows a Background Image block in the Content tab, with a preview, Replace image and an Image URL box. On an image element, uploading a video in the Content tab's Upload Media zone sets the element's video, and uploading an image clears it.
#The Media panel
The Media item in the icon rail is a chat-style AI image and video generator with a permanent Library that follows your account across devices.
| Setting | Options |
|---|---|
| Image models | GPT Image 2 (the default), Nano Banana Pro, Nano Banana 2, Seedream 5, FLUX 2 Pro, FLUX Schnell |
| Video models | Kling 3.0, Seedance 2.5, Veo 3.1, Seedance 2.0, Luma Ray 2, Wan 2.1 |
| Aspect ratios | 16:9, 1:1, 9:16, 4:3 |
| Library tabs | All, Images, Videos, Favorites and Trash, with a search box |
- You can favourite an item, delete it to Trash and restore it, download it, copy it and re-run it.
- Each AI generation costs credits. The price depends on the model.
#AI chat and media
The AI chat can generate an image, generate a video, and set or remove a video on an element. Those generations are charged by the image or video model, on top of the chat message. Attaching an image to a chat message makes that turn an escalated one, which costs 30 credits. See Edit with AI chat.
#Limits
- Fresh AI images on a generated site are made for signed-in builds. Guest builds use curated images only.
- A newly added image element starts with a placeholder photo. Replace it before you publish.
- Your site's media library as seen in your dashboard is described in Media library.
#Common questions
Images over 10 MB and videos over 50 MB are refused. Other types than those listed are not accepted. In the homepage prompt box HEIC files are rejected, so export them as JPG.
The element you generated for gets the image at once. The Media panel's Library keeps what you generate there. The Images tab's Recent generations list is cleared when you close the editor.
Use Object fit in the Images tab: Cover, Contain, Fill or None.
Use the Media panel, or ask the AI chat to generate a video for the element.