Weekend project: vector DB benchmarking

I built a small prototype around speech 006.

What it does

• Solves a tiny but annoying problem
• Focused on speed + clarity

Feedback I want

• UX flow
• Edge cases
• What to add/remove