Wan 2.7
Generate and edit videos with Wan 2.7 using text, frames, references, or an existing clip.
Loading inspiration gallery.
What Is Wan 2.7?
Wan 2.7 is an AI video model built for both generation and revision. It can create a new clip from a prompt, animate an opening image, connect a first frame to a final frame, continue a short source video, or use image and video references to guide a new scene.
Its editing workflow starts from an existing video and follows a written instruction, with an optional reference image for added direction. Text and image workflows support 2–15 second output at 720p or 1080p, while reference and editing limits adapt to the source material. The practical advantage is choice: begin with whatever part of the shot is already decided, then use the matching workflow to control what happens next.
What Can You Create with Wan 2.7?
Turn a Prompt into a Shot or Short Sequence
Start with a written scene when no source footage exists yet. Describe the subject, action, setting, camera movement, lighting, and shot order, then choose a 2–15 second duration, 720p or 1080p output, and landscape, portrait, square, 4:3, or 3:4 framing. Wan 2.7 interprets single-shot or multi-shot direction from the prompt, so a clear timeline matters more than a list of disconnected visual ideas.
Control How Motion Starts—and Where It Lands
Use one image when the opening composition should anchor the clip, or add a final frame when both ends of the motion are already approved. First-and-last-frame guidance is useful for reveals, transitions, and story beats that need a defined destination. Wan 2.7 can also continue from a short video segment, giving you a way to extend an existing action instead of inventing the entire shot from text.
Guide Subjects with Image and Video References
Use reference images or videos when a person, character, object, movement, or visual role needs to remain recognizable in a new scene. Name each reference in the prompt and explain exactly what it contributes. Keeping one clear role per source makes the request easier to interpret and the result easier to review than uploading several assets without assigning their purpose.
Revise a Strong Clip Instead of Starting Over
Start from an existing video and describe the change you want to see. Wan 2.7 can follow instructions for a focused visual, motion, narrative, camera, or style revision, and an optional reference image can clarify the new direction. Keep the edit narrow enough to compare against the source, especially when identity, timing, composition, or audio continuity should remain recognizable.
How to Use Wan 2.7
Choose the Right Starting Point
Select Text for a new idea, Image for an approved opening frame, Frames for a defined start and finish, Reference for subject-led guidance, or Video Edit when a source clip already contains the performance or composition you want to keep.
Describe the Shot and Assign Every Input
Write one coherent direction covering subject, action, setting, camera, lighting, and timeline, then explain the role of every uploaded frame, reference, source video, or audio track instead of leaving the relationship implicit.
Set the Available Controls and Review
Choose the duration, resolution, framing, and workflow-specific options, generate the clip, then compare motion, composition, reference consistency, edit scope, and audio against the original brief before deciding whether to refine it.
Best-Fit Workflows for Wan 2.7
Plan a Product Reveal Between Approved Frames
Use an approved hero image as the first frame, or add a final packshot when the reveal must land on a precise composition. Wan 2.7 is a useful choice here because frame guidance defines the visual boundaries while the prompt concentrates on camera movement, product action, and the transition between them.
Carry a Character or Object into a New Setting
Use separate image or video references to establish the subject, prop, movement, or voice role, then write a new scene that explains how those elements interact. Wan 2.7 fits this task because its reference workflow is designed to interpret multiple source roles rather than relying on the prompt to recreate every defining detail from memory.
Make a Focused New Version of an Existing Clip
Use Video Edit when the source already has a useful performance, camera path, or story beat but one part needs to change. Wan 2.7 lets the brief focus on the revision—such as style, setting, movement, or emphasis—so you can review what changed without discarding everything that already works.
Wan 2.7 Prompt and Input Best Practices
Describe one ordered action with a clear camera path
Avoid unrelated events competing inside one short clip
Give every reference one named and visible role
Avoid crowded sources and ambiguous role assignments
Request one change you can compare against the source
Avoid changing style, story, camera, and identity at once
Wan 2.7 vs Wan 2.6 vs Seedance 2
Best for
Moving between new generation and focused revision without changing to a different model family.
Established short-video recipes built from text, one opening image, or one to three reference videos.
Reference-heavy projects that need mixed image and video guidance or a wider output matrix.
Starting material
Text, one opening image, first and last frames, a short source clip, image or video references, or an existing video to edit.
A prompt, exactly one opening image, or one to three short reference videos.
Text, one opening image, first and last frames, or a larger collection of image and video references.
Workflow range
Text to video, image to video, first-and-last-frame video, video continuation, reference to video, and video editing.
Text to video, image to video, and reference to video, without Wan 2.7's separate continuation and editing paths.
Text to video, image to video, multi-reference video, and start-and-end-frame video.
Clip controls
720p or 1080p; text and image workflows support 2–15 seconds, while reference and editing limits depend on the source media.
720p or 1080p; 5, 10, or 15 seconds for text and image, and 5 or 10 seconds for reference video.
Auto or 4–15 seconds, with 480p, 720p, 1080p, and 4K choices available across its exposed workflows.
Audio approach
Text and image workflows can use a supplied audio track; editing can keep source audio or let the model handle audio automatically.
Text and image modes accept a supplied WAV or MP3 background track with clear trim or trailing-silence behavior.
Uses a generated-audio option rather than Wan 2.7's supplied-audio and source-audio editing choices.
Choose it when
You want the broadest Wan workflow for moving from prompt or frames to references, continuation, and revision.
Your existing workflow already fits its simpler input modes and does not need video continuation or instruction-based editing.
Mixed visual references, first-and-last-frame direction, or a broader resolution range matters more than video editing.
Choose Wan 2.7 when you want one model to cover prompt-led generation, frame control, reference guidance, video continuation, and focused editing. Choose Wan 2.6 for a simpler established Wan recipe, or Seedance 2 when broader reference capacity and output choices are the deciding factors. None is a universal quality winner—the best choice depends on the source material and the kind of control the shot needs.
FAQs about Wan 2.7
What is Wan 2.7?
Wan 2.7 is an AI video model for generation, reference-led creation, continuation, and instruction-based editing. It can begin from text, frames, references, or an existing clip, which makes it useful when different stages of a video idea already have different amounts of visual direction.
What can I create with the Wan 2.7 AI video generator?
You can create a new video from text, animate an opening frame, guide motion between first and last frames, continue a short video, build a new scene from image or video references, or revise an existing clip. The inputs and controls change with the workflow you choose.
Does Wan 2.7 support first-and-last-frame video?
Yes. Add an opening image and a final image when you want to define both ends of the motion. This is useful for planned reveals, transitions, or short narrative beats where the destination matters as much as the starting composition.
Can Wan 2.7 use image and video references?
Yes. Its reference workflow can use images and videos to guide subjects, objects, movement, and other visual roles in a new scene. Name the references in the prompt and explain what each one should contribute instead of expecting the model to infer every relationship.
Can Wan 2.7 edit an existing video?
Yes. Start with one source video, describe the visual, motion, narrative, camera, or style change, and add a reference image when it helps clarify the target. Focused instructions are easier to compare against the source than a request that changes every part of the clip at once.
How long can Wan 2.7 videos be?
Text and image workflows support integer durations from 2 to 15 seconds. Reference and editing limits depend on the source-media rules, so the available duration can be narrower when a source video is involved.
Which resolutions and aspect ratios does Wan 2.7 support?
Wan 2.7 supports 720p and 1080p output. Prompt-led and reference workflows expose multiple framing choices, while image-led and editing workflows can inherit more of their composition from the source material.
Can Wan 2.7 generate or use audio?
Text and image workflows can use a supplied WAV or MP3 track, and text-to-video can create matching background audio when no track is provided. Video editing can preserve the source audio or use its automatic audio setting, depending on the chosen options.
Is Wan 2.7 free to use?
This guide is free to browse. Generating or editing a video uses Credits, and the current quote is shown before you submit the request.