Scheduled Maintenance

We're performing scheduled maintenance. Some features may be temporarily unavailable.

Choose Theme

JollyAI Default

Clean and professional dark theme

It Comes at Night New

Ultra dark JollyAI theme

Stranger Things

80s horror-sci-fi aesthetic

Batman

Dark knight, dark theme

Barbie

Bright and colorful Barbie theme

Ocean Blue

Twitter-inspired blue theme

Midnight Purple

Catppuccin-inspired purple theme

Forest Green

Matrix-style green theme

Crimson Red

Warm red-based dark theme

Amber Glow

Warm amber/orange theme

Dracula

Popular gothic-inspired theme

Monokai

Classic Sublime Text theme

Nord

Arctic-inspired cool theme

Gruvbox

Community favorite warm theme

Solarized Dark

Carefully calibrated for readability

One Dark Pro

Popular VS Code theme

How to Make an AI Video from an Image — Free, No Account

You have a photo or an illustration, and you want it to move. Here's the complete guide — from uploading your image to downloading your finished AI video in minutes.

Try Image-to-Video Now — Free

What This Guide Covers

What is AI Image-to-Video?

AI image-to-video is exactly what it sounds like: you give the model a static image and it generates a short video clip based on it. The AI doesn't just pan over your photo — it actually synthesises new frames, inferring how objects in the scene would move, how light would behave, and what motion would make sense given the content.

The result is a 3–5 second video clip that feels like your image came to life. A portrait photo might have subtle wind movement in the hair. A landscape might have rippling water and drifting clouds. A product shot might gently rotate or have light playing across it.

Until recently this kind of generation cost money or required serious hardware. JollyAI runs Wan 2.2 and LTX-2 on cloud servers and gives you access completely free, in your browser.

Step-by-Step: Animate an Image on JollyAI

1

Go to the Image-to-Video Tool

Open jollyai.online and select the Image-to-Video workflow. No account, no login — just the tool. If you want to use LTX-2 specifically, go to LTX-2 Image-to-Video.

2

Upload Your Image

Click the upload area and select your image — JPG or PNG both work. The image should be reasonably clear and well-composed. Blurry, heavily compressed, or very small images will give worse results. Aim for at least 512x512 pixels.

3

Write Your Motion Prompt

This is the part most people underestimate. The prompt tells the AI how to animate the image. Don't just write what's in the picture — describe the motion you want. "Camera slowly panning right" is a better prompt than "a forest". See the prompt guide below for full details.

4

Click Generate

Hit generate and wait. Wan 2.2 takes roughly 3–5 minutes. LTX-2 is faster at 1–2 minutes. Both run on cloud servers — you don't need to keep any tab open or leave your computer on during this time.

5

Download Your Video

When generation is complete, your video appears in the output. Click download — it's an MP4 file with no watermark. Use it however you want: post it, embed it, edit it further in any video editor.

That's it. No sign-up screen, no credit check, no watermark. Five steps and you have your video.

Wan 2.2 or LTX-2 — Which Should You Use for Your Image?

Both models support image-to-video, but they produce different kinds of motion. The choice comes down to what kind of image you have and what kind of movement you want.

Wan 2.2 — better for:

  • Portraits and people — natural movement, hair, clothing
  • Nature photos — water, trees, clouds, animals
  • Cinematic or atmospheric images
  • Action and physical motion
  • Any image where realistic physics matters
Use Wan 2.2 →

LTX-2 — better for:

  • Product photos — subtle motion, lighting shift
  • Artwork and illustrations — stylised animation
  • Images where sharp detail should be preserved
  • Abstract visuals
  • When you need the result faster
Use LTX-2 →

Not sure? Generate the same image with the same prompt on both and compare. Since both are free, there's no reason not to try both.

What Images Work Best?

Not every image will produce an impressive result — the quality of your input has a big effect on the output. Here's what helps:

Images that tend to work well:

  • Clear, sharp photos with a defined subject
  • Images with natural scenes (sky, water, vegetation, fire) — easy to animate
  • Portraits with visible facial features and clear lighting
  • Product shots with clean backgrounds
  • Illustrated characters or artwork with clear subject boundaries
  • Landscapes and architectural photos with depth

Images that often produce weaker results:

  • Very small images (under 256x256px)
  • Heavily compressed JPEGs with artefacts
  • Images with complex text or logos
  • Screenshots with UI elements
  • Extremely dark or overexposed images
  • Cluttered scenes with many tiny objects

Quick tip: If your image isn't producing good results, try running it through the JollyAI Image Upscaler first to improve resolution and clarity, then use it as input for image-to-video.

Writing the Motion Prompt — The Part Most People Get Wrong

This is where most people stumble. They upload a photo of a forest and type "a beautiful forest" as their prompt. Then they're confused when the video just looks like their photo is zooming slightly.

The model already knows what's in your image. The prompt's job is to tell it what should happen — what moves, how it moves, where the camera goes.

The formula for a good image-to-video prompt:

[camera movement] + [subject motion] + [atmospheric detail] + [style/mood]

Camera movement vocabulary:

What you want Words to use
Camera moves toward subject "slow zoom in", "dolly in", "push in"
Camera pulls back "pull back", "zoom out", "dolly back"
Camera rotates horizontally "pan left", "pan right", "horizontal pan"
Camera rotates vertically "tilt up", "tilt down", "crane shot"
Camera follows subject "tracking shot", "follow cam"
Camera stays still, scene moves "static camera", "locked off"
Camera floats or drifts "floating camera", "gentle drift"

Subject motion vocabulary:

Subject type Motion words
Hair / clothing "gently blowing in the wind", "flowing"
Water "rippling", "flowing gently", "waves lapping"
Trees / plants "leaves rustling", "branches swaying"
Sky / clouds "clouds drifting slowly", "stars twinkling"
Fire / smoke "flames flickering", "smoke rising"
Portrait / person "subtle breathing", "slight head turn", "eyes blinking slowly"

Real Prompt Examples That Work

Here are proven prompt structures for common types of images. Copy and adapt these directly.

Portrait or headshot photo

"Subtle movement, hair gently blowing in a light breeze, soft blinking, slight head turn, cinematic, shallow depth of field"

Works with: Wan 2.2. Result: a portrait that breathes and feels alive.

Landscape or nature photo

"Slow cinematic pull-back, clouds drifting across the sky, water rippling gently in the foreground, leaves rustling, golden hour light"

Works with: Wan 2.2. Result: an epic landscape that feels like a film opening shot.

Product photo

"Gentle slow 360-degree rotation, studio lighting sweeping from left to right, product catching the light, sharp and detailed"

Works with: LTX-2. Result: a polished product showcase video.

Digital artwork or illustration

"Magical particles floating around the character, cape flowing, dramatic camera tilt upward, cinematic atmosphere, ambient light pulsing"

Works with: Wan 2.2. Result: illustration comes to life like a game cinematic.

Fire, water or atmospheric image

"Flames flickering and dancing, embers floating upward, smoke rising, static locked-off camera, warm orange glow"

Works with: Wan 2.2. Result: mesmerising ambient loop perfect for social media.

Architecture or interior

"Slow floating camera drift through the space, light shifting naturally, subtle dust particles in the air, cinematic wide shot"

Works with: LTX-2 for sharp interiors, Wan 2.2 for atmospheric mood.

When It Doesn't Look Right — Quick Fixes

Problem: The video barely moves — it just looks like a slideshow

Fix: Your prompt isn't describing enough motion. Add explicit camera movement ("slow zoom in") and subject motion ("hair gently blowing in breeze"). Even one or two clear motion phrases make a big difference.

Problem: The video looks blurry or low quality

Fix: Check your source image — if it's small or low-quality, upscale it first. Also try LTX-2 if you were using Wan 2.2 — LTX-2 tends to preserve fine detail better.

Problem: The subject changes or deforms badly

Fix: Your prompt is asking for too much motion. Scale it back — "subtle" and "gentle" in your prompt signal the model to keep changes minimal. For portraits especially, less motion instruction = better identity preservation.

Problem: The motion looks unnatural or jerky

Fix: Add "smooth", "fluid", "slow", or "cinematic" to your prompt. Also check that your image has clear depth — flat, two-dimensional-looking images sometimes produce less natural motion.

Problem: The output is completely different from my image

Fix: If your prompt includes a lot of descriptive text about what's in the image, the model may be treating it as a text-to-video prompt and partly ignoring the image. Keep the prompt focused on motion, not description. Let the image speak for itself.

Bonus: Character Animation with Wan Animate

Image-to-video is great for general scene animation. But if you have a character image and want to give it specific, controlled movement — a walking cycle, a dance, a fighting pose sequence — there's a more powerful option: Wan 2.2 Animate.

Wan Animate takes two inputs: your character image, plus a reference video showing the motion you want. The model reads the pose and movement from the reference video and applies it to your character. The result is your character performing the exact motion from the reference video.

This is particularly useful for:

  • Animating your own character designs or OC (original character) art
  • Making portraits dance or perform gestures
  • Creating consistent character animation across multiple clips
  • Game character concept animation
Try Wan 2.2 Animate — free →

Common Questions

Is this really free with no watermark?

Yes. JollyAI is free — no account, no payment, no watermark. Every video you download is clean.

How long can the output video be?

JollyAI generates clips up to around 5 seconds in length. This is typical for current AI video models and is enough for most social media and B-roll use cases. For longer content, generate multiple clips and edit them together.

What image file formats are supported?

JPG and PNG both work. Keep your image under 10MB for fastest upload. For best results, use a clear image with at least 512x512 pixels.

Can I use my phone to generate videos?

Yes. JollyAI works in mobile browsers. You can upload an image from your phone's camera roll and generate directly. The process is the same — no app needed.

Can I use the generated video commercially?

Wan 2.2 is Apache 2.0 licensed, permitting commercial use. For LTX-2, check Lightricks' licence terms for commercial projects. For personal and creative use, both are unrestricted.

What's the difference between image-to-video and Wan Animate?

Image-to-video generates free-form motion from your image guided by a text prompt — the AI decides how to move things. Wan Animate uses a reference video to give the model explicit motion to copy, which is more controlled but requires a suitable reference video.

Ready? Animate Your Image Now

Pick your model, upload your image, write your motion prompt, and generate. It takes 5 minutes.

Wan 2.2 Image-to-Video — Free LTX-2 Image-to-Video — Free

Related Guides and Tools