AI Integrator makes pictures and short video clips from a written description or from a photo you already have. You can work in the Images studio, where you see the price of each try before you press Generate, or ask Chat, one of your Agents, or Code to make one for you. Every result is saved to Files.
How do I make an image with AI?
Open Images in the Chat sidebar, choose a model, describe the picture, and press Generate. The studio has 38 image models, including GPT Image, Nano Banana, Flux, Grok Imagine, Seedream, Recraft, Qwen, Krea, MAI Image and Muse, and it shows only the shapes, sizes and quality settings the chosen model accepts.
Each studio session keeps up to 200 tries. Pick an earlier one to put it back on stage, and your next description builds on it. To keep a face, a product or a style the same across pictures, add references: an upload, a picture from Files, a photo from your camera, or an earlier try. New image starts a fresh session.
Enhance makes the picture on stage larger and sharper with Topaz, doubling each side up to about 4K. It saves the result as a new try and leaves the original as it was.
How do I turn a photo into a video?
Switch the studio's model menu from Image to Video, choose a model and a clip length, and press Generate. With a picture on stage, the clip starts from it; without one, it is made from your description alone. A clip takes a minute or two, and you can leave while it is made: it lands in the studio and in Files, under Generated, then Video.
| Model | What it makes | Length |
|---|---|---|
| H3 Max Turbo | Fast, lower-cost clips from a description or a picture | 5 to 15 seconds |
| H3 Max | The strongest H3 clips, with up to four reference pictures | 5 to 15 seconds |
| FLUX 3 Max | Clips with sound, from a description or a picture | Up to 20 seconds |
All three run on Fal, which hosts the models.
Can I extend a video, recast it, or lip-sync a face?
Yes. Each of these works on the clip or picture on stage and saves the result as a new clip:
- Extend adds 3 to 15 seconds. Quality continues the motion and sound. Turbo starts from the last frame for about half the price, so the motion can change where the clips join.
- Recast puts the people in up to four of your photos into a video of up to 30 seconds. Your description can say who replaces whom.
- Lip Sync makes the face in a picture speak or sing an audio clip of 5 to 60 seconds. You can record the audio in the + menu or pick it from Files.
How much does it cost?
Images and video use your plan's usage, like a Chat message. The studio's composer shows what the next try will take before you press Generate. It reads About when the final amount depends on what the model returns. Each try holds that usage before it starts and settles after, and a clip that fails or that you cancel gives it back. Pricing lists the plans.
Can Chat, Agents and Code make videos?
Yes. Turn on Image Generation or Video Generation in Skills & Plugins, then ask in any chat: "Make a short clip of this photo with the camera pushing in slowly." Chat, your Agents and Code use the same models and the same saved files as the studio, and the clip appears in the conversation when it is ready.
Video costs more than text or pictures, so a clip that could take more than $1 of your usage waits for you. The model tells you the price and asks, and only your reply can approve it.
- You
- Chat
- Video model
- You to Chat: Make a 15-second clip of this photo
- Chat to You: This clip costs more than $1. Make it?
- You to Chat: Yes
- Chat to Video model: Starts the clip
- Video model to Chat: Saved to Files
- Chat to You: Shows the clip
Where are my images and videos saved?
In AI Integrator Files, under Generated. From there you can use them as references for the next picture, attach them to a chat, or share them. The Images and video page lists every model and edit in one place.
