We find that large-scale web datasets often repeat the same objects in diverse poses and scenes. Leveraging this, we create a massive supervised dataset for subject-driven generation and object insertion. Our method achieves state-of-the-art identity preservation without requiring test-time fine-tuning.