Glass Bloom
3:4 · 8 s · 2KFirst frameMiniMax H3
AI image to video generator Bring a still image to life
Upload a photo, product shot or illustration, describe how it should move, and Felo turns it into a video clip. The agent plans the camera, chooses a video model and keeps the context so you can refine the result in the same thread.
- Input formats
- JPG, PNG, WebP, HEIC/HEIF
- Max image size
- 30 MB
- Length
- 2–30 s
- Resolution
- 480P–2K
Image to video model directory
Felo routes each shot to a video model. MiniMax H3, Seedance 2.5, Wan 3.0 Prime and Kling 4.0 are available now and all accept image inputs; Gemini Omni Flash is coming soon. Pick one in video settings, or let the agent decide.
ModelStatusBest forImage inputLengthMax resolution
MiniMax H3
Available nowAvailable now2K clips from text, image, video and audio references
Up to 9 reference images
4–30 s
2K
Image input: Up to 9 reference imagesLength: 4–30 sMax resolution: 2K
Seedance 2.5
Available nowAvailable nowFast, motion-heavy short clips
First & last frame, up to 30 references
4–30 s
720P
Image input: First & last frame, up to 30 referencesLength: 4–30 sMax resolution: 720P
Wan 3.0 Prime
Available nowAvailable nowFlexible generation modes
First & last frame
2–30 s
1080P
Image input: First & last frameLength: 2–30 sMax resolution: 1080P
Kling 4.0
Available nowAvailable nowCinematic motion and character consistency
First, middle & last frame, up to 10 references
3–30 s
1080P
Image input: First, middle & last frame, up to 10 referencesLength: 3–30 sMax resolution: 1080P
Gemini Omni Flash
Coming soonComing soonFast multimodal generation
—
—
—
Image input: —Length: —Max resolution: —
Each model has its own upload limits, frame options and aspect ratios. The composer shows the options for the model you pick.
What is image to video AI?
Image to video AI animates a still image. You provide the picture and describe what should move — a person turning, fabric lifting, a camera pushing in — and the model generates new frames that continue from it. Because the first frame is fixed, you keep control of the subject, composition and style instead of describing them from scratch.
- Start from a photo, product shot, illustration or generated image.
- Describe the motion and camera; the subject stays recognizable.
- Choose 21:9, 16:9, 4:3, 1:1, 3:4 or 9:16.
- Refine with a follow-up in the same thread.
Made with Kling 4.0
What you can make with image to video
Two examples for each job, each made with Kling 4.0 from one still image. Drag across a clip to compare the image with the video.
Start from a packshot and name the label and shape that must stay fixed. Let light, liquid or props move around the product.
Your imageKling 4.0 videoRosemary Serum1:1 · 5 s · 1080P
Prompt: Keep the bottle shape and position fixed. Water droplets slowly roll down the glass, a beam of light sweeps across the slate from left to right, and the camera makes a slow, smooth half-orbit around the bottle.
Your imageKling 4.0 videoMorning Pour16:9 · 5 s · 1080P
Prompt: Hot black coffee is poured into the mug from above, steam curls up through the sunbeam, the linen curtain moves slightly in a breeze, and the camera pushes in slowly toward the mug.
Your imageKling 4.0 videoRooftop Organza9:16 · 5 s · 1080P
Prompt: A strong gust of wind lifts and ripples the organza gown, she turns her head toward the camera and takes one step forward, city lights twinkle behind her, slow push-in. Keep her face and the dress consistent.
Your imageKling 4.0 videoGolden Hour Portrait9:16 · 5 s · 1080P
Prompt: She blinks slowly, turns her face into the sunlight with a small smile, a few strands of hair move in a light breeze and the sunlight flickers softly across her skin. Keep her face identical.
Your imageKling 4.0 videoNight Diner16:9 · 5 s · 1080P
Prompt: Rain grows heavier on the window, the neon flickers twice, the headlights of a passing car sweep across the booths, and the camera slowly dollies along the window.
Your imageKling 4.0 videoSunflower Time-Lapse9:16 · 5 s · 1080P
Prompt: Turn the photo into a time-lapse: clouds race across the sky, wind sends waves through the sunflowers, and the light shifts from afternoon gold to a deep orange sunset while the tractor stays still.
How to turn an image into video in Felo
Three steps from a still image to a finished clip.
- 01
Choose your image
Upload a JPG, PNG, WebP, HEIC or HEIF image up to 30 MB. A sharp, well-lit picture with a clear subject gives the model the most to work with.
- 02
Describe the motion
Say what should move and how: a head turn, fabric in the wind, a slow push-in. Also say what must stay unchanged, such as the face, product label or background.
- 03
Generate, refine and download
Felo picks a video model and matches the aspect ratio, resolution and duration to your image. Preview the clip, ask for a change in the same thread, then download or share the version you like.
How to write an image to video prompt
The image supplies the look; the prompt supplies the motion. Cover these five points and the agent can plan the shot.
- 01
Starting image
Describe what the image contains and what must stay fixed: the subject, framing, colors and any text or product label.
- 02
Subject motion
What moves, in which direction and how fast. One main action per shot reads more naturally.
- 03
Camera move
Push in, pull out, track left, arc around or hold static. Small, slow moves work well for portraits and products.
- 04
What stays unchanged
Name the face, outfit, geometry or label that must not change, and anything that should stay out of frame.
- 05
End state
How the shot should finish. With first-and-last-frame generation, the second image defines the final frame.
Begin exactly from the uploaded image and keep the woman's face, profile and the pearl-gray background unchanged. A single silver serum droplet lands on her cheekbone and slides slowly downward. The camera trucks right with a small, slow move while she turns five degrees toward the light. The droplet breaks into fine micro-droplets that stay on the skin. Natural blink, no beauty filter, no text, no logo.
Image to video examples made with Felo
Every clip below was generated by animating a still image — some from a single first frame, others from first and last frames. Watch one, then start from your own image.
Mirror Skin
3:4 · 6 s · 768PFirst frameMiniMax H3
Citrus Orbit
1:1 · 6 s · 768PFirst frameMiniMax H3
Folded Wind
3:4 · 8 s · 768PFirst & last frameMiniMax H3
Two Stops Apart
16:9 · 9 s · 768PFirst frameMiniMax H3
The Ocean Above the Crossing
9:16 · 8 s · 2KFirst & last frameMiniMax H3
Related video tools
Pick the page that matches what you are making.
Key takeaways
- Felo image to video animates a still image into a clip of 2 to 30 seconds, depending on the model.
- Seedance 2.5 and Wan 3.0 Prime take a first and last frame, Kling 4.0 adds a middle frame, and MiniMax H3 works from up to 9 reference images.
- Upload images up to 30 MB each; accepted formats and how many you can add depend on the model.
- Six aspect ratios cover cinematic 21:9, widescreen 16:9 and vertical 9:16.
Frequently asked questions
- What is an image to video AI generator?
- An image to video AI generator turns a still picture into a video clip. Felo lets you upload an image, describe the motion and camera, then generates a clip that continues from that image. The agent keeps the context so you can refine the result in the same thread.
- How do I turn a photo into a video with Felo?
- Upload your photo in the Felo video agent, describe how it should move, and press generate. Felo plans the shot, picks a video model and returns a clip you can preview, refine or download.
- Which image formats and sizes can I upload?
- It depends on the model. MiniMax H3 accepts JPG, PNG, WebP, HEIC and HEIF images up to 30 MB, 256–5,760 px per side, with up to 9 images per request. Kling 4.0 accepts JPG and PNG up to 30 MB, at least 300 px per side, with up to 10 images. Video settings check the limits of the model you pick.
- What should I put in an image to video prompt?
- Describe what moves, how the camera moves, and what must stay unchanged, such as the face, product label or background. One main action per shot works best. For first-and-last-frame generation, describe how the scene travels from the first image to the second.
- Which video model does Felo use for image to video?
- MiniMax H3, Seedance 2.5, Wan 3.0 Prime and Kling 4.0 are available now and all accept image inputs; Gemini Omni Flash is coming soon. Seedance 2.5 and Wan 3.0 Prime can use images as the first and last frames, and Kling 4.0 also supports a middle frame. The agent picks a model for each shot, or you can choose one in video settings.
- How long can a Felo image to video clip be?
- Clip length depends on the model: Wan 3.0 Prime runs 2–30 seconds, Kling 4.0 3–30 seconds, Seedance 2.5 4–30 seconds, and MiniMax H3 4–30 seconds (4–15 seconds at 1080P and 2K). You can pick the duration in video settings before generating.
- What aspect ratios and resolution can I use?
- Felo supports 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16; each model offers a subset. Resolution runs from 480P to 2K: MiniMax H3 goes up to 2K, Wan 3.0 Prime and Kling 4.0 up to 1080P, and Seedance 2.5 up to 720P. The output aspect ratio can differ from the source image, so choose a ratio that fits where you plan to publish.
- Can I control the first and last frame?
- Yes. Seedance 2.5 and Wan 3.0 Prime take a first and a last frame, and Kling 4.0 also takes a middle frame. Choose the model in video settings, upload the images and describe the motion that connects them. MiniMax H3 has no frame slots; with it, the agent writes into the prompt which reference image the clip should start and end on.
- What is the difference between image to video and text to video?
- Image to video starts from a picture, so the subject, composition and style are already set; your prompt only describes the motion. Text to video builds every frame from a written description. Felo supports both in the same thread.
- How much does Felo image to video cost?
- Image and video generations use Felo points. The cost depends on the model, resolution and length; video settings show an estimate before you generate, and the final charge may vary. The pricing page lists the points included with each plan.
Bring your image to life
Upload a photo, product shot or illustration and describe how it should move.