Image + prompt → video

One picture and a prompt become a moving shot.

What it takes

• A starting picture — it becomes frame one, exactly as it is
• A prompt for what happens next

What it gives back

• An MP4, played inline and downloadable

What actually runs

• Wan 2.2 I2V-A14B: two 14-billion-parameter experts, one that decides the motion and one that finishes the picture.
• Both carry a lightx2v 4-step LoRA, which is the difference between four sampler passes and twenty — most of what a job costs.
• The high-noise expert samples first and hands its leftover noise to the low-noise one halfway through.

Worth knowing. Describe physics, not mood. “The camera pushes in slowly while the box turns a quarter turn” lands; “epic cinematic vibes” does not.

Running it

The .json attached to this listing is the graph itself — the one the Forge runs, not a tidied-up copy. Values in {{double braces}} are the placeholders filled in per job; replace them with real values and it runs in any ComfyUI new enough for the nodes it uses.

Full setup — the model manifest, the directories, the traps — is at /workflows/video_i2v.

Licence and credit

The graph is ours and you may use it for anything, including commercially. The models it loads are not ours: each carries its own licence from whoever trained it, and that is the one that governs what you may do with the output.