Wan 2.1 VACE Outpaint · Video to Video

ABOUT THE WORKFLOW

Expand a Video's Frame
Upload a video. The workflow pads the frame with empty space, masks the new areas, and Wan 2.1 with the VACE module fills them in by reading the lighting, depth, motion, and content from the original footage. The generated content moves and animates consistently with the source video. The output includes the expanded video and a side-by-side comparison with the original.

Model

• Wan 2.1 T2V 14B by Alibaba with the VACE (Video-conditioned All-in-one Controllable Editing) module. VACE conditions the generation on the existing video frames and masks, so the expanded areas match the original scene's motion, perspective, and lighting across every frame.

HOW IT WORKS

Step 1. Upload your video
The video you want to expand. The workflow loads up to 97 frames and resizes to 854x480 before padding.
Works great with: live-action footage · drone shots · stock video · AI-generated clips

Step 2. Set the padding direction and amount
Choose how many pixels to add on each side (left, top, right, bottom). The default adds 128px of white space. The white areas become the regions the model fills in.

Step 3. Write your prompt
Describe what the expanded areas should contain. "Soldiers gathered in a forest" tells the model what scene to continue. Match the prompt to what is already in the video.

Step 4. Hit run
Wan 2.1 with VACE fills the padded areas frame by frame, matching the motion and lighting of the original content.

Step 5. Download
The workflow outputs the expanded video and a side-by-side comparison at 24fps as MP4.
Ready for: Premiere · DaVinci Resolve · After Effects · social media · broadcast

First time? Upload a video, set the padding to 128px on the sides you want to expand, write a prompt describing the scene, and hit run.

RECOMMENDED SETTINGS

Quick-start guide. Find the goal that matches yours and copy the settings.

• Standard video expansion — 128px padding, 20 steps, CFG 4, shift 8, seed randomized. Upload your video, write a scene prompt, run.

• Expand horizontally to widescreen — Set left and right padding to 128px each. Top and bottom stay at 0. Converts a narrow or square video to a wider aspect ratio.

• Expand vertically for portrait — Set top and bottom padding to 128px each. Left and right stay at 0. Extends the frame upward and downward.

• More padding for a bigger expansion — Increase padding beyond 128px. Larger expansions give the model more area to fill but take longer and may produce softer edges.

• Filled areas do not match the original — Rewrite the prompt to describe the scene more specifically. "Dense pine forest, overcast light, soldiers in camouflage moving through undergrowth" gives the model more context than a generic description.

• Motion in the expanded area looks inconsistent — VACE reads motion from the original frames. If the source video has fast or unpredictable movement at the edges, the expanded area may struggle to follow. Use source clips with gradual edge motion for the smoothest results.

Prompt: Describe the scene that should appear in the expanded areas. "Forest clearing with tall pine trees, dappled sunlight, fallen leaves" is specific. "Nature scene" is not. The prompt fills in what the model cannot infer from the existing frames.

LEARN

📹 Videos

• Intro to Floyo

• ComfyUI 101 Free Course ft. Sebastian Kamph

• Floyo 101 for Team Collaboration

✨ Quick links

• Help docs

• LoRA Training guide

• Discover more workflows

USE CASES

🖥️ Widescreen Conversion
Turn a square or vertical video into a widescreen format by expanding the left and right edges with matching scene content.

🎬 Reframing for Post-Production
Add breathing room around a tightly cropped shot for titles, lower thirds, or re-composition without reshooting.

📱 Aspect Ratio Adaptation
Expand a 16:9 clip to 9:16 for vertical platforms, or a 1:1 square to landscape, by filling the added space with consistent scene content.

🎮 Environment Extension for Game and VFX
Expand the frame of a reference clip to reveal more of the environment for previsualization, matte painting reference, or set extension planning.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works great

• Videos with clear environmental context at the edges (sky, ground, walls, foliage)

• Evenly lit scenes without heavy vignetting or lens distortion

• Moderate expansion (128px per side at 480p base)

• Prompts that match the visible scene content

⚠️ May produce softer results

• Videos with subjects cropped tightly at the edge (half a face, half an object)

• Very fast motion or camera shake at the frame boundary

• Massive expansion (more than 256px per side in a single pass)

• Prompts that describe content unrelated to the existing scene

FAQ

What is VACE in Wan 2.1?
VACE stands for Video-conditioned All-in-one Controllable Editing. It is a module that attaches to the Wan 2.1 video model and conditions generation on existing video frames plus spatial masks. For outpainting, it reads the original video and fills masked (padded) areas with new content that matches the scene's motion, lighting, and perspective frame by frame.

How does video outpainting differ from image outpainting?
Image outpainting extends a single still image. Video outpainting extends every frame of a video while maintaining temporal consistency. The expanded areas move, light, and animate in sync with the original footage across the full clip, not frame by frame in isolation.

Can I expand a video in all four directions at once?
Yes. Set padding values for left, right, top, and bottom simultaneously. The model fills every direction at once. Results work best when the original video has clear environmental cues on all edges.

What resolution does the output have?
The workflow resizes the input to 854x480 before padding. The final output dimensions depend on how much padding you add. At 128px per side horizontally, the output is approximately 1110x480. The video saves at 24fps as an H264 MP4.

Does video outpainting preserve the original audio?
The current workflow does not pass through audio. If you need the original audio, recombine it with the expanded video in an editor.

Is Wan 2.1 VACE licensed for commercial use?
Wan 2.1 is an open-source model by Alibaba. The VACE module is released alongside it. Check the current license terms on the model page for commercial use in your specific project.

How to run video outpainting with VACE online?
You can run video outpainting with VACE online through Floyo. No installation, no setup, no local GPU needed. Open the workflow in your browser, upload your video, and hit run. Free to try.

WHY FLOYO?

Floyo is the only platform with team collaboration for ComfyUI in the browser. You run workflows with no install. You share run history, assets, and models across your team. You pay only when you generate. Floyo supports open-source and closed-source models.

A designer runs an edit and likes the result. A teammate opens that exact run from shared history and keeps going. No file handoffs. No version confusion.

For studios and enterprise teams, Floyo adds private workspaces, pooled resources, and a team usage dashboard. Other ComfyUI cloud tools run for one person at a time. Floyo runs for the whole team, with transparent per-generation costs.

Ready to try it?
Upload a video, set the padding, describe the scene, and hit run.

→ Launch Workflow, Free

Questions? Watch the free course or check the FAQ above.