How do you recover after realizing your OSS v0.1.0 was wrong?

I recently shipped v0.1.0 of a small open source project and quickly realized that while it "worked" the architecture and scope were fundamentally wrong.

Instead of stacking fixes, I decided to step back and rebuild it properly toward v1.0.0, with a much stronger focus on simplicity and contributor-friendliness.

I’m curious how others here have handled similar situations:

- When you realize early architectural decisions were wrong, how far back do you roll things? - How do you balance “learning by building” with not over-engineering? - At what point does it make sense to invite contributors after a reset?

I’m less interested in validation and more in learning how OSS maintainers think about resets like this.