Ads with accurate branding
Logos, packaging, and brand colors come through accurately: for commercial ads, product clips, and e-commerce where recognizability matters.
Video generation with Wan 2.7
Video generation with Wan 2.7
Flami lets you create videos with the Wan models from Alibaba: Wan 2.7, Wan 2.6, Wan 2.5, Wan 2.2 Fast, and Wan 2.2. Four modes in one model: create from a description, from an image, from a reference, or edit an existing video.
Why Wan 2.7
A model from Alibaba with a unique combination of four modes: creating from text, an image, a reference, and editing a finished video. Thinking mode plans the shot before generation.
Four modes in one model
One service covers the whole video workflow: creating from a text description, from a source image, from a reference that replicates voice and motion, and editing an already-finished video from a description.
Thinking mode before generating
Wan 2.7 thinks through the prompt first: it analyzes the composition, the frame's logic, and camera movement, then starts generating. Fewer artifacts and closer alignment with your idea.
Up to 9 reference images in one request
In I2V mode, you can submit a grid of nine images as references for complex scenes. The model uses them to control the composition, angles, and look of the shot.
Editing a finished video from a description
The unique Video Edit mode: change objects, style, or color, or restore old footage, all with one text instruction, no regenerating from scratch.
What Wan 2.7 can do
Duration
2-10 seconds per clip
Quality
720p, 1080p
Formats
16:9, 9:16, 1:1, 4:3, 3:4
Inputs
Text plus up to 5 image or video references, plus optional audio
Sound
Native with lip-sync, or submit your own audio source to drive it
Modes
Creation from text, from an image, from a reference, or editing a finished video
Creation methods
From a description, from a source image, from a start and end frame, or matching the voice from an audio reference
Wan 2.7 video for any task
Narrative clips with distinct characters
Different character faces with no AI same-face effect: for indie film, episodic content, and campaigns with a large cast.
Long narratives up to 15 seconds
Full ad stories, multi-shot narratives, and extended scenes in a single pass, with no cuts.
Editing archival video
Colorizing black-and-white footage, restoring old material, swapping objects, and restyling: all with one text instruction applied to an already-shot clip.
Complex scenes with a 9-image reference grid
Multi-angle compositions built from a grid of nine reference images: for interiors, multi-angle portraits, and complex cinematic composition.
How it works
Step 1. Write a prompt
Describe the video you want: the scene, mood, camera movement, characters. You can specify HEX codes for brand colors and text that should appear in the frame.
Step 2. Add an image
Upload up to 9 reference images to set the style, character, or scene. You can attach a start and end frame for precise control over how the clip opens and closes.
Step 3. Get your finished video
Click Generate, and our service will create your Wan 2.7 video with the settings you chose. Thinking mode plans the shot first, then starts generating, which takes a few minutes.
Create your first video with Wan 2.7
One workspace, all the world's best video models