Mischief 1 Image to Video: How to Animate Any Picture

Mischief 1 image to video is the only way Mischief 1 works: you give it one still picture as the first frame, describe the motion in words, and it returns a short, silent widescreen clip. This guide walks through the process step by step, explains what this image to video model does with your image, and shares practical tips for choosing first frames that animate well. If you just want to try it, use the generator right below – upload a still, describe the motion and render a clip without leaving this page.

What happens when you turn an image into video

Mischief 1 is an image-to-video model released by The Biological Computing Co. (TBC) on October 30, 2026, and built on Alibaba's open Wan 2.2 model (OpenRouter, The Next Web). Your picture fixes the subject, composition, colours and lighting. Your prompt decides what moves. The model cannot work from text alone: Treza Labs reports that a request without an image is rejected with a 400 error before anything renders (Treza Labs).

In Treza's test, it kept the still's composition and added the prompted action around it. That is the right mental model: you are not asking for a new scene, you are asking the existing scene to move for a few seconds.

Output you can expect

fal's schema notes that clips run very slightly longer than the requested length because of how frames are counted (3 s becomes 3.0625 s, 5 s becomes 5.0625 s). Because output is 16 fps, very fast motion can look choppier than on a 24 fps model, as Treza points out.

Step by step: Mischief 1 image to video on this site

1. Open the generator. Go to the generator on the homepage and sign in. New accounts get free credits as listed on the pricing page.
2. Upload your first frame. JPG, PNG or WebP, up to 4 MB. A landscape image works best, because Mischief 1 always outputs widescreen (see the cropping tip below).
3. Describe the motion. Say what moves, how fast, and what the camera does. Also say what should stay still. Our Mischief 1 prompt guide has examples you can copy.
4. Pick resolution and length. 480p or 720p, 3 or 5 seconds. The button shows the credit cost before you generate (1 credit per second at 480p, 2 per second at 720p).
5. Generate and download. When the clip is ready you can preview it and download the MP4. If a generation fails, the credits are refunded automatically.

How to choose a first frame that animates well

Treza Labs sums it up well: because every clip opens on your image, the quality of the image is the quality of the shot. Some practical guidelines:

Shoot or crop to widescreen

fal fits your image to the render size with a center crop by default, so a vertical phone photo loses its top and bottom. Crop to landscape yourself before uploading so you decide what stays in frame.

Keep the subject clear

A single, well-lit subject with some space around it gives the generator room to add motion – a slow push-in, drifting steam, hair moving in wind. Cluttered frames with many small objects give it more chances to distort details.

Leave room for the motion you want

If you want a car to drive out of frame, compose it with road ahead of it. If you want a camera to pull back, start fairly close. The image to video generator has to invent anything that is not in the picture.

Use a sharp, well-exposed image

Blur, heavy compression and noise in the source carry into the video. Render at least at the resolution you need: fal notes images are never scaled below the requested resolution.

Generated images work too

If you only have an idea, not a photo, generate a still with any image model first, then animate it here. Treza describes this two-step "make the frame first" workflow as the way to use it from text.

Prompt tips specific to image to video

• Describe change, not content. The AI video generator already sees the picture. Spend words on motion and camera, not on re-describing what is in the frame.
• Be specific and fairly long. fal's schema says long, specific prompts steer the underlying WAN model best.
• Name the camera move. "Slow push-in", "static shot", "gentle pan left". The Wan 2.2 guide on fal recommends "motion description + camera movement" for image-to-video.
• Lock what should not move. "The mug, table and background stay exactly as they are."
• Do not ask for sound. The output has no audio track.

A simple example for a product photo of a coffee cup:

Static camera. Thin steam rises from the coffee and drifts slowly to the right. The cup, saucer, table and background stay perfectly still. Soft morning light, no cuts.

Known limits (and what we don't know yet)

• No text-only mode, no vertical or square output, no clips longer than 5 seconds, no sound.
• Only the first frame can be set – no last frame or extra reference images (Treza Labs).
• TBC says it is about five times faster than Wan 2.2 at the same quality. No independent benchmark had been published at the time of writing (TNW, OrigoTech).
• We have not published our own quality tests, so we make no quality claims beyond what the sources say.

Want to know what a batch of clips costs? See Mischief 1 pricing. Curious how it differs from its base model? Read Mischief 1 vs Wan 2.2.