FLUX 3 Max image to video turns a still picture into a 5 to 20 second video with sound: your image becomes the first frame, and the prompt describes what moves from there. Use it when the look of the shot has to stay exactly as in your picture, such as a product, a character or an illustration.
Updated October 21, 2026.
Five steps in the browser, about a minute end to end.
Open the editor on Flux3Max and choose Image to video. Sign in with Google; new accounts get free credits for one 5-second 720p video.
A JPG, PNG or WebP up to 10 MB and 4096 × 4096 pixels. It becomes the first frame of the video exactly as it is, so crop and fix it before you upload.
Write what happens next, not what the picture shows: the one main action, how the camera moves, and the sounds you want to hear.
Any length from 5 to 20 seconds, 720p or an upscaled 1080p, and sound on or off. Leave the aspect ratio on Auto to keep the shape of your picture.
The clip appears next to the editor and in My videos. Credits are only taken when the video is delivered; if it fails, they go back.
fal describes it the same way: the image sets the opening composition, and the prompt describes what changes after the first frame.[fal image-to-video API] The same photo of a woman by a window, with two prompts:
Before
A beautiful woman in a red dress standing by a window, cinematic, 4K, highly detailed, masterpiece.
After
She turns from the window toward the camera and smiles as a breeze lifts the curtain beside her. Slow push-in from a medium shot to a close-up, soft late-afternoon light. Sound: distant street traffic below, the curtain rustling, a quiet laugh.
What the first prompt misses
For FLUX 3, fal notes that quiet, low-motion scenes are the most likely to come back with little audible sound, and that several coordinated actions at once are harder than one clear action.[fal]
A starting picture you can take with your phone, and the prompt that animates it.
Your picture
A phone selfie of a person holding your product (here, a closed burger box) in a kitchen.
Prompt · 8 s · 9:16
Authentic UGC selfie video, phone front camera, warm kitchen light, handheld. The young woman from the photo holds up the burger box to the camera, opens it, lifts the burger with the cheese stretching and takes a big bite, then grins and nods. Sound: room tone, the cardboard lid opening, a crunchy bite, and her saying “okay, this is so good” with her mouth half full.
Why it works: Starting from your own photo keeps the product and the person exactly as they are; the prompt only adds the action and the line.
The first frame is your image as it is, so a new outfit, background or face can only appear as something that happens during the clip. If you want a different starting look, change the picture first.
Crop the picture to the shape you will post in, such as 9:16 for Shorts or 16:9 for YouTube, and leave the aspect ratio on Auto, so the video keeps the framing you chose.
One clear action holds up better than three things happening together, as fal notes for FLUX 3.[fal] Make a second clip for the second action.
An image-to-video prompt has a second job, as fal’s FLUX 3 guide puts it: saying what to leave alone, such as the label on the product, the face or the background.[fal FLUX 3 guide]
Upload two pictures and the clip starts on the first and ends on the second; FLUX 3 Max fills in the motion between them. Use it for transitions, before-and-after shots and product reveals that have to land on a set frame. Describe the path between the two in the prompt.
Start from an image when the look must not change: your product, your character, your artwork. Start from text when you are still exploring an idea. Both modes rank near the top: on October 21, 2026, FLUX 3 Max was #2 in Artificial Analysis’ image-to-video leaderboard (1,138) and #4 in text to video (1,159).[Artificial Analysis image to video][Artificial Analysis text to video]
Open the editor on Flux3Max, choose Image to video and upload a JPG, PNG or WebP. Describe the motion, the camera and the sound, pick a length from 5 to 20 seconds, and generate. Your image becomes the first frame of the clip.
It is the first frame, so the video starts on your picture exactly. From there the model animates it according to the prompt; anything you ask to change happens on screen during the clip.
With Auto, the video follows the shape of your image. You can also choose 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 or 2:1; crop the picture to that shape first so the framing is yours.
No. On Flux3Max a video costs the same credits per second in every mode; only the length and the resolution change the price.
Yes, with First + last frame: upload a second picture and the clip ends on it, with the model filling in the motion between the two. Use it for transitions, before-and-after shots and product reveals that have to land on a set frame.
Usually because the prompt describes the picture instead of an action. Name one thing that happens, how the camera moves and what we should hear; a prompt with no action leaves little to animate.
Figures from model makers are their own measurements unless stated.