If Wan made an image editor, wouldn't character consistency be solved?

The Wan 2.2 model, released over a year ago, is gaining renewed attention for its ability to maintain consistent character appearance across different images—a persistent challenge in AI image generation. At higher resolutions, Wan 2.2 delivers near-perfect likeness preservation, a feat that current tools like Qwen IE, Flux Klein, and Kontext struggle to match, even with custom LoRAs (low-rank adapters). Users note that Wan's low-noise variant can also be used for direct image generation, though it remains finicky to set up.

A Wan-based image editor could revolutionize character consistency by leveraging its animation framework, FFGO, which already excels at preserving likeness across multiple reference images and styles. The community would likely create custom LoRAs for style transfer overnight, making the tool highly flexible. However, Wan's shift away from open-source development suggests they may not prioritize this, leaving a gap for competitors to fill.