What it does
autostar turns a goal you can measure into a structured optimization loop. You define what a good result looks like, and an evaluator scores each attempt. The agent runs experiments, learns from every attempt and works toward the best version within the budget you approve.
Use cases
- 01Improving a prompt against a fixed scoring rule
- 02Comparing variants of a page and keeping the best one
- 03Tuning a process with a clear measure of success