Wan 2.2 SVI four-pass V3

Most AI video models stop at about 5 seconds. This Wan 2.2 video extension workflow (SVI) chains four 5-second segments into one continuous video of up to 20 seconds.

Great for:

• Chaining several actions into one story
• Longer AI influencer and character videos
• Cartoon scenes that continue naturally

How it works

1. Load your start image.
2. Write a prompt for each of the four segments. Each segment can also use its own LoRAs.
3. Run it. The segments are blended together, upscaled and interpolated to 60 fps.

Tip for higher quality

Remove the turbo LoRA from the high noise sampler, raise CFG to 2.2-3 and add a few more steps. Generation gets slower, but motion and detail improve.

Make it yours with LoRAs

This workflow supports LoRAs. Grab character, style, anime or cartoon LoRAs from CivitAI or Hugging Face, drop them into the LoRA loader, and keep your AI character looking the same in every shot.