I built a logic engine simulator that generates sequence diagrams, variable networks, and formula flows directly from natural language input.

It's a visual tool for reasoning: users can input statements like "gravity(earth → object)" and see how variables propagate, even without predefined formulas.

This is the "safe version" — no core algorithm or AI logic is exposed, but it's fully interactive and designed to simulate how real reasoning works.

I'd love feedback on usability, potential use cases, and how this compares to LLM-based approaches.

Demo and code in the repo!

[Before learning(image)] https://i.postimg.cc/7Y8rdPFq/Figure-1-2025-07-06-3-54-57.png

[See the Brain Simulate Learning] https://youtu.be/5wTX7lzmPog

[After learning(image)] https://i.postimg.cc/mk7JBMvB/Figure-1-2025-07-06-3-56-34.png

Unlike LLMs that rely on massive training data, this system builds logical pathways from structured reasoning.