Sidecar improves prompt-image alignment and character consistency across SDXL and FLUX-based models without requiring additional training or architectural modifications

Experiments on FreeStoryBench show that Sidecar consistently improves prompt-image alignment and character consistency across multiple SDXL- and FLUX-based baselines, with negligible computational overhead.

http://arxiv.org/abs/2608.27280v1