After seeing DeepMind's AlphaEvolve [1], I wanted to try building something similar. This isn't the first open-source attempt [2] [3] [4], and it won't be the last, but I focused on making mine more modular and customizable to make experimenting with different evolution-based approaches or prompting strategies easier (e.g. balancing exploration and exploitation).
The vision is to make it easy to swap components and experiment, though there's still work to do.