Made with this app
How it works
Wan 2.7 is Arteza's text-to-video app built on the latest Wan model from Alibaba, designed for creators who need genuinely smooth motion and faithful scene reproduction in short-form video clips. It accepts both plain text prompts and reference images as input, producing high-quality clips between 5 and 10 seconds. Whether you are animating a product concept, visualizing a narrative scene, or turning a still image into motion, Wan 2.7 is built for work where motion fluidity and visual accuracy are the priority.
How to use this app
- 1
Describe what you want to create, or upload your source file.
- 2
Pick your options, then press Generate.
- 3
Watch your result appear in the gallery within moments.
- 4
Download, share, or generate again with a new idea.
What you can make
Product Visualization in Motion
Designers and marketers can describe or photograph a product and let Wan 2.7 animate it with smooth, artifact-free motion. The model's enhanced motion smoothness keeps rotating objects, pouring liquids, or unfolding materials looking physically plausible across the full clip duration.
Animating Still Reference Images
Because Wan 2.7 accepts image input alongside text, you can drop in a finished illustration, concept painting, or photograph and describe the motion you want. The model preserves the visual character of the source image while adding coherent, fluid movement frame by frame.
Narrative Scene Construction
Screenwriters, storyboard artists, and indie filmmakers can write a scene description and receive a high-fidelity clip that captures lighting, environment, and character movement. Scene fidelity means background details and spatial relationships stay consistent throughout the generated clip.
Social Content with Complex Motion
For creators who need motion-heavy social clips, such as camera pans, crowd movement, or weather effects, Wan 2.7's enhanced motion capabilities handle multi-element scenes without the stuttering or ghosting that affects simpler models, making the output more usable with minimal editing.
Iterative Creative Exploration
Because both text and image inputs are supported, creators can rapidly iterate: start with a text prompt, pull a frame from the result as an image input, then refine with an adjusted description. This loop lets you home in on a precise visual and motion outcome efficiently.
Prompt ideas to try
- A glass bottle of perfume rotates slowly on a white marble surface, soft studio lighting, subtle smoke curling around the base, photorealistic.
- A narrow cobblestone street in a rainy European city at night, reflections shimmering on the wet stones, a single figure walking away with an umbrella.
- Aerial view of a dense forest canopy during golden hour, a slow forward camera drift revealing a hidden lake below, cinematic color grade.
- A ceramic coffee mug sits on a wooden desk while steam rises gently and morning light shifts across the wall behind it, warm interior atmosphere.
- Using the attached concept illustration as reference, animate the waterfall on the left side of the frame with natural flowing motion, keeping the surrounding cliffs static.
- A close-up of a mechanical watch face, second hand sweeping smoothly, macro lens depth of field, polished metal surfaces catching diffused light.
Why creators use this app
- Enhanced motion
- Scene fidelity
- Text + Image input
- Latest model
Tips for better results
Describe Motion Explicitly
Wan 2.7's enhanced motion engine responds well to direct motion language. Instead of describing only the subject, specify camera behavior, object movement direction, and speed. Phrases like 'slow dolly forward' or 'gentle clockwise rotation' produce more controlled results than leaving motion implied.
Use Image Input as an Anchor
When scene fidelity matters most, supply a reference image alongside your text prompt. The model uses it to lock colors, spatial layout, and visual style, reducing the chance of unwanted variation between what you envisioned and what the model generates.
Specify Clip Duration Intent
Wan 2.7 supports 5 to 10 second clips. If your scene has a single sustained motion, a shorter duration keeps it tight and avoids filler frames. For scenes with a camera move plus a subject action, lean toward 8 to 10 seconds so both elements have room to read clearly.
Ground Your Scene With Environment Details
Scene fidelity improves when your prompt includes concrete environmental context: surface materials, light source direction, time of day, and atmosphere. Vague settings give the model less to lock onto, which can result in backgrounds that drift or feel disconnected from the foreground subject.
When to choose this app
Choose Wan 2.7 when motion smoothness and scene fidelity are non-negotiable. If you have compared results from Seedance 2.0 Mini for quick low-cost drafts or LTX-2 Pro for its own distinct rendering style, Wan 2.7 sits in the tier that prioritizes fluid, artifact-free movement and consistent visual environments across the clip. Its image input support also makes it more versatile than text-only generation workflows.
Frequently asked questions
Can I use my own photograph as the input image?
Yes. Wan 2.7 accepts image input directly alongside your text prompt. You can upload a photograph, illustration, or rendered concept image, then describe the motion or environmental change you want applied. The model attempts to preserve the visual identity of your source image while adding the specified motion.
What clip lengths can Wan 2.7 produce?
Wan 2.7 generates clips between 5 and 10 seconds. You can orient your prompt toward the shorter end for tight, focused motion sequences or toward the longer end when your scene involves multiple distinct movements or a camera journey that needs time to resolve.
How does scene fidelity differ from basic video generation?
Scene fidelity refers to the model's ability to keep background environments, spatial relationships, lighting consistency, and fine details stable across frames. In lower-fidelity models, backgrounds can morph, textures can flicker, and objects can shift position unexpectedly. Wan 2.7 is specifically tuned to reduce these inconsistencies.
Does Wan 2.7 handle fast action or is it better suited to slow motion?
Wan 2.7's enhanced motion smoothness benefits both fast and slow action, but it is particularly strong where inter-frame consistency matters: fluid dynamics, rotating objects, crowd movement, and camera pans. Very rapid action with extreme frame-to-frame change is inherently harder for any model to render without artifacts.
What happens if I provide an image input but no text prompt?
Providing a text prompt alongside your image is strongly recommended. Without a text description, the model has less directional information about what motion or scene change to generate, which can produce unpredictable results. Even a short description of the intended motion meaningfully improves output quality and relevance.
Can Wan 2.7 generate video with realistic human movement?
Wan 2.7 can generate clips featuring human subjects in motion. As with all current text-to-video models, complex articulated human motion such as detailed hand gestures or rapid athletic sequences benefits from clear, specific prompting. Describing posture, direction, and pace gives the model more structure to work with.
Which AI model powers this app?
This app runs on Wan 2.7, available through Arteza with no separate account or setup.
Can I use the results commercially?
Yes. Content you generate is yours to use, subject to our content licenses.
How long does a generation take?
Most generations finish in under a minute, and you can watch progress live in the gallery.