Cinematic Establishing Shot
One slow forward dolly over a golden-hour city: neon boardwalks, glass towers, and waves held steady for 8 seconds.
Alibaba’s Wan 2.7 video model in four modes: write a prompt, start and end on images you choose, guide a scene with reference images and a clip, or upload a short video and edit it with words. Clips run up to 15 seconds in 720p or 1080p.
Vibrant cinematic establishing shot drifting forward over a futuristic coastal city at golden hour, neon-lined boardwalks, glass towers reflecting amber light, distant ocean waves crashing, warm sunset color grade, the camera moves at a very slow forward dolly, hyper-detailed 8k
Reference example. It is not a result generated from your upload.
Wan 2.7 (often typed wan2.7 or wan 27) is a video generation model from Alibaba’s Wan team, the group behind the Wan AI family that people also know as WanX AI. It follows Wan 2.6 in the lineup and brings four ways to create into one model: text to video from a prompt, image to video that opens on a first frame and can land on a last frame, reference to video guided by images and a short clip, and video editing that rewrites a clip you upload.
Wan 2.7 renders in 720p or 1080p, with clips of up to 15 seconds from text or images and up to 10 seconds in reference mode, in 16:9, 9:16, 1:1, 4:3, or 3:4. It reads camera language literally, so one clear move per shot and named shots for longer clips give clean results. Looking for stills instead? Wan 2.7 Image (wan2.7-image) has its own page.
Wan 2.7 video clips with the prompts behind them. Press a button to load a similar idea into the workspace.
One slow forward dolly over a golden-hour city: neon boardwalks, glass towers, and waves held steady for 8 seconds.
Fog, glassy water, and a single boat in 1080p, with a warm highlight on one peak to break the cool palette.
"Shot 1… Shot 2… Shot 3…": the sprint, the leap, and the landing come out as clean, separate beats.
A held frame that slowly pushes forward, letting the light and fog do the moving.
Sky to shoreline in one slow tilt, the kind of shot you can pin down with a first and last frame.
A slow pan across red brick toward the light: ideal source footage for a season or weather edit.
Three steps in your browser, with no local install or GPU.
Text for a scene from words, Image to open on a first frame (and optionally end on a last frame), Reference for up to three images plus an optional clip, or Edit video to change a 2–10 second clip you upload.
Describe the subject, action, one camera move, and the light. In Reference mode, name your files as Image 1, Image 2, and Video 1; in Edit video, say what changes and what stays the same.
Choose 720p or 1080p, 5, 10, or 15 seconds, and an aspect ratio. The cost appears on the Generate button before you start; download the Wan 2.7 video when it is done.
Wan 2.7 image to video opens on the exact first frame you upload, such as a product photo, a storyboard panel, or a still you own, and can finish on a last frame you choose. The model fills in the motion between them, which is ideal for transitions, reveals, and loops that must land on a precise image.
Upload a 2 to 10 second clip and describe the change: a new season, different weather, another outfit, or a restyled look. Wan 2.7 video edit keeps the motion and the framing and returns a clip with the same length and shape as your source.
Add up to three reference images of characters, products, or places and an optional reference clip, then call them Image 1, Image 2, Image 3, and Video 1 in the prompt. Wan 2.7 keeps those faces and objects consistent in a brand-new scene of up to 10 seconds.
Text and image clips run 5, 10, or 15 seconds, long enough for atmospheric scenes and true multi-shot sequences. Draft in 720p, then render the keeper in 1080p, where fog, water, foliage, and material textures hold up best.
Which Wan model to pick on EzImgMaker.
| Wan 2.7 | Wan 3.0 | Wan 2.2 | |
|---|---|---|---|
| Best for | Frame control and editing your own clips | Longer scenes with native audio | Quick, low-cost 5-second drafts |
| Text to video | Yes | Yes | Yes |
| Image to video | First frame, or first + last frame | First frame, or first + last frame | One starting image |
| References | Up to 3 images + 1 clip here | Images, videos, and audio | No |
| Edit an existing video | Yes (2–10 s clips) | No | No |
| Video length | 5, 10, or 15 s (reference: up to 10 s) | 2–30 s | About 5 s |
| Resolution | 720p, 1080p | 480P, 720P, 1080P | 480p or 720p |
| Aspect ratio | 16:9, 9:16, 1:1, 4:3, 3:4 | Auto plus five ratios | 16:9 or 9:16 |
Image to video follows the shape of your first frame, and an edit keeps the shape and length of your source clip.
A quick Wan 2.7 prompting guide: subject + action + one camera move + light, and name your shots for long clips.
Only an idea? Text. A strong still? Image. Characters or products to keep? Reference. Footage you like? Edit video.
"Slow forward dolly", "crane up", "90-degree rotational pan": Wan reads these literally, and stacking moves confuses it.
For 10 to 15 seconds, write "Shot 1: … Shot 2: … Shot 3: …" so the cuts come out clean.
With a first frame, spend your words on movement and camera; the image already sets the subject and style.
In edits and reference clips, end with "keep the people, motion, and camera the same" or "keep her raincoat unchanged".
Fog, water, and product materials reward 1080p; test motion and action ideas in 720p first.
Previsualize shots between two key frames and test multi-shot sequences.
Product spins, reveals, and square or vertical ads from a product photo.
Change the season, weather, or style of a short clip without reshooting it.
9:16 clips and recurring characters built from a few reference images.
Other Wan versions and video generators on EzImgMaker.
The image side of Wan 2.7: create and edit stills, including start frames for this page.
Read the guideAlibaba’s newest Wan model: up to 30 seconds with native audio.
Read the guideFast, low-cost 5-second clips and speech-to-video portraits.
Read the guideAlibaba’s film-style model with sound and lip-synced speech.
Read the guideAnimate a photo with the model that suits it best.
Read the guideTurn a written idea into a clip.
Read the guideCommon questions about the Wan 2.7 AI video model.
Wan 2.7 is an AI video model from Alibaba’s Wan team. It makes clips of up to 15 seconds in 720p or 1080p from text, from a first frame (optionally with a last frame), or from reference images and a clip, and it can edit a short video you upload.
Wan 2.7 is free to try on EzImgMaker: new accounts get free credits when you sign up. After that, each clip uses credits based on its length and resolution, and the exact cost is shown on the Generate button before you start.
Text and image clips can be 5, 10, or 15 seconds. Reference clips go up to 10 seconds. An edited clip keeps the length of your 2 to 10 second source video.
Yes. Choose Image, upload the photo as the first frame, and describe the motion and camera. The video opens on your photo and keeps its shape. Only use photos you own or have permission to use.
The first frame is the exact image the clip opens on; the optional last frame is the image it should land on. Wan 2.7 fills in the movement between them, so describe what happens in between, such as "the café fills with customers".
Yes. Choose Edit video, upload an MP4 or MOV of 2 to 10 seconds (up to 100 MB), and describe the change, such as a new season, weather, outfit, or style. The motion and framing stay, and the result keeps the clip’s length.
Use Wan 2.7 to edit your own clips or to pin a shot between a first and a last frame. Use Wan 3.0 for longer scenes of up to 30 seconds with native audio and richer references.
720p and 1080p. Text and reference clips can be 16:9, 9:16, 1:1, 4:3, or 3:4; image to video follows your first frame and edits follow your source clip.
Describe the subject, one action, one camera move, and the light in a sentence or two. For longer clips, split the action into "Shot 1, Shot 2, Shot 3". With a first frame, describe the motion rather than the picture. See the Wan 2.7 prompting tips above.
Wan is developed by Alibaba’s Wan team (the model family was earlier branded Tongyi Wanxiang, which is why some people search for WanX AI). EzImgMaker is an independent site and is not affiliated with, endorsed by, or sponsored by Alibaba.
Write a prompt, start from your own frames, or upload a clip to edit, and let Wan 2.7 render it in 1080p.
Wan, Alibaba, and Alibaba Cloud are trademarks of Alibaba Group Holding Limited or its affiliates. EzImgMaker is an independent site and is not affiliated with, endorsed by, or sponsored by Alibaba or the Wan team. Model names are used only to describe compatibility.