Skip to main content
Nuzza
Upgrade
Loading account

Wan 2.7

Wan 2.7 brings text, opening-frame, first-and-last-frame, reference-to-video, continuation, and instruction-based editing into one model family. Nuzza exposes the connected workflows with operation-specific controls for duration, resolution, framing, references, audio, and source media.

Describe what you want to create with Wan 2.7
Text, images, references, and source video feeding separate Wan 2.7 workflows

Choose the Workflow That Matches Your Source

Start from text, one opening frame, matched first and last frames, reference images or videos, or an existing source clip. Choosing the correct operation first keeps the prompt focused on the transformation Wan 2.7 is actually being asked to perform.

Try For Free
A coherent opening frame, endpoint, and video continuation connected on one timeline

Control an Opening Frame, Endpoint, or Continuation

Anchor the first composition with an image, add a compatible end frame when the destination matters, or continue from a short source clip. Each path solves a different continuity problem, so review identity, geometry, lighting, and camera direction across the handoff.

Try For Free
Recurring characters and products remaining recognizable across several cinematic shots

Keep Referenced Subjects Recognizable

Use reference images or videos to define recurring people, objects, wardrobe, movement, or visual treatment. Assign each source a clear role in the prompt and check every shot for identity, proportion, color, and relationship drift.

Try For Free
An existing daytime video transformed into a controlled evening version while preserving the subject

Edit Existing Video with Instructions

Upload a source clip and describe the requested visual change or style transfer instead of rebuilding the scene from zero. Wan 2.7 can preserve source structure while applying an edit, but targeted objects, edges, motion, and untouched regions still require review.

Try For Free
Dialogue, ambience, music, and effects aligned with a multi-shot video sequence

Plan Video and Audio Together

Supported generation modes can automatically create matching sound or use supplied audio, while editing can preserve source audio or let the model handle it. Describe dialogue, ambience, music, and effects beside the visible event they should support, then review synchronization and balance.

Try For Free
One polished short-form sequence prepared in landscape, square, and portrait HD formats

Build Short HD Sequences up to 15 Seconds

Official text- and image-led Wan 2.7 workflows support 720p or 1080p output and clips from 2 to 15 seconds. Reference and editing operations have their own shorter limits, so the selected operation—not the family headline—determines the valid request.

Try For Free

What Is Wan 2.7 Built For?

A coherent multi-shot product campaign with stable subjects and matching sound

Multi-shot campaign stories. Create a connected product or brand sequence with planned shot changes, recurring subjects, controlled pacing, and matching sound in one short-form deliverable.

The same character and product preserved across several new scenes

Character and product continuity. Use reference media to carry recognizable people, wardrobe, props, or products through new settings and camera angles while checking every output for drift.

A source clip continuing into a visually consistent new story beat

Video continuation and endpoints. Extend a short clip or connect compatible first and last frames when a sequence needs a controlled handoff, reveal, or additional beat.

A focused video revision changing the environment while preserving the performance

Instruction-led video revisions. Relight, restyle, or revise selected content in an existing clip while keeping useful motion, timing, composition, and source audio where the workflow allows.

How to Use Wan 2.7 on Nuzza AI For Free

A reference image added to an AI creation interface beside the submit control
Step 1

Choose the Correct Wan 2.7 Workflow

Start with text, an opening image, first and last frames, references, a continuation clip, or a source video for editing.

A structured prompt being written in an AI creation interface
Step 2

Assign Every Input a Clear Role

Describe the subject, action, camera, light, sound, and intended change, then set the controls exposed for that specific operation.

A completed AI-generated video ready to review in the result interface
Step 3

Generate and Inspect the Complete Result

Review continuity, motion, edits, frame edges, reference fidelity, dialogue, ambience, and audio synchronization before downloading or revising.

Wan 2.7 vs Wan 2.6 vs Seedance 2

Feature / Model Wan 2.7 Wan 2.6 Seedance 2
Best Fit Projects that move among frames, references, continuation, generation, and revision. Established text, opening-image, or reference-video recipes. Projects centered on mixed visual, video, and audio reference direction.
Starting Inputs Text, opening frame, first and last frames, references, audio, or source video. Text, one opening image, or a small set of reference videos. Text, frames, images, video, and audio across its connected modes.
Workflow Range Generation, reference-to-video, continuation, and instruction-based editing. Generation and reference-led video without Wan 2.7's connected editor. Generation and multimodal reference workflows rather than Wan-style editing.
Duration Up to 15 seconds for text and image generation; limits vary for references and editing. Up to 15 seconds in the longer text and image paths. Up to 15 seconds in the current Nuzza mapping.
Audio Automatic or supplied audio in supported generation modes; source-audio handling in editing. Supplied or generated audio behavior depends on the operation. Generated-audio control is available in supported workflows.

FAQs about Wan 2.7

What is Wan 2.7?

Wan 2.7 is Alibaba's multimodal video family for text- and frame-led generation, reference-to-video, continuation, audiovisual creation, and instruction-based video editing.

Is Wan 2.7 free to use?

Yes. Nuzza offers free Credits for new users to try Wan 2.7. Once your free Credits are used up, you will need to subscribe to a paid plan to continue.

What can I use as input?

Depending on the selected operation, Nuzza accepts text, one opening image, first and last frames, reference images or videos, optional audio, a continuation clip, or a source video for editing.

How long can Wan 2.7 videos be?

Official text- and image-led workflows support 2–15 seconds. Reference-to-video, continuation, and editing have separate limits, so check the visible controls for the selected operation.

Can Wan 2.7 generate or use audio?

Yes. Supported workflows can create matching audio or use a supplied track, while video editing can preserve source audio or let the model handle it automatically.

Can Wan 2.7 edit an existing video?

Yes. The current editing workflow accepts a source clip and an instruction, with optional references and audio-handling controls where exposed.

What resolution does Wan 2.7 support?

Official connected workflows use 720p or 1080p output, with availability depending on the operation.

How should I use references?

Use clean, relevant sources, assign each one a clear identity or style role, and state what must stay recognizable. References guide the model but do not guarantee perfect preservation.

What should I review before publishing?

Check character and product identity, physical motion, shot continuity, edits, crop, text, dialogue, sound balance, and audiovisual synchronization across the entire clip.

Move from a Prompt or Source into a Finished Video

Choose the Wan 2.7 workflow that matches your real starting material, assign every reference a role, and review the complete audiovisual result.

Try Wan 2.7 for Free Now