Video · wan-video/wan-2.7-i2v
WAN 2.7 I2V — frame into living video
WAN 2.7 I2V takes a still image and produces a video with coherent motion, automatically picking matching audio. That is its main edge over the younger WAN 2.5 I2V Fast: here you get a synced soundtrack out of the box, plus first-and-last-frame control and clip continuation.
Where it earns its keep
- I2V with audio out of the box — the model picks audio for the scene or accepts yours.
- First-and-last-frame: pin the start and the end, the model builds the transition.
- Clip continuation: hand in a short fragment, the model extends the motion.
WAN 2.7 I2V vs WAN 2.5 I2V Fast
WAN 2.5 I2V Fast is about speed and cost. If the job is to animate a frame quickly for social, with no audio and no frame-level control, it is cheaper and good enough.
WAN 2.7 I2V earns its keep when audio, 5–10 second length, and movement accuracy matter. Worth it for ad-grade dynamic clips with background music; overkill for a one-off “make it move”.
What you need on input
A first frame in jpg/png/webp up to 20 MB. Optionally: a last frame for ending control, audio (wav/mp3, 3–30 seconds), or a video clip (mp4/mov 2–10 seconds) for continuation.
Resolutions 720p and 1080p, length 5–10 seconds. The cost depends on length and resolution and is shown in the app.
Frequently asked about WAN 2.7 I2V
Does the model make the audio?
Yes — if you do not pass audio, the model generates something matching the scene. If you want your own track or voiceover, upload it and the motion syncs to it.
How is first-and-last-frame different from plain I2V?
You pin both key frames and the model builds the transition. Useful for templated clips — a logo at the start and the end, an explicit close to the scene.
How many seconds can I generate?
5 to 10 seconds per run. For longer videos use clip continuation: feed in a short fragment and the model extends from there.
What image formats are accepted?
JPG, PNG, BMP, WEBP, up to 20 MB. Cleaner frames produce cleaner motion — noisy input shows up as artifacts.