Evogine – Zero-dependency evolutionary optimization for Python
Six optimizers, zero dependencies, agent-steerable mid-run—genuinely thoughtful design for research.

CMA-ES optimization finds real stars matching your text—Neal Agarwal's version doesn't do this.
Curious users, educators, and anyone who enjoys interactive visualizations
Neal Agarwal's Constellation Draw · SkyMap · Stellarium
I was inspired by Neal Agarwal's constellation-draw, but approached it as an inverse fitting problem. Given text, find real stars that trace the letters. It uses a coarse grid search plus CMA-ES refinement to search for matches. You can reroll results, switch to a globe view, and drag vertices around to tune the constellation manually.
Example for the input "HI HN": https://starspelled.com/#BAEAAKAVRDAKF3___7uEBFOfsQQjABlKPmu...
Source: https://github.com/JoshuaMosier/written-in-the-stars
Feedback welcome!
Six optimizers, zero dependencies, agent-steerable mid-run—genuinely thoughtful design for research.
Yet another code-to-LLM formatter when repomix and git2llm already exist.
Another daily word game clone with user-generated content, but the builder looks rough.
Memory-mapped files enable CPU training on terabyte datasets without GPU.
Markdown-to-mind-map rendering is slick, but Obsidian Canvas already maps local notes better.
Schema-driven Markdown registry beats drawing tools for living documentation.