A demo that explores a Neo4j graph toward a goal. At every node, the ways out are offered to Jev as a multiple-choice question, and Jev returns how likely each one is to lead somewhere useful. In the same call it answers whether the goal has been reached.
Several promising routes can be followed at once, and the best path is shown in an interactive graph view alongside the nodes around it.
Built as a Jupyter notebook by Michael Hunger.
What you get
- Jupyter notebook source, MIT licensed
- Interactive graph view of each run
Jev
Python
OpenRouter
FastAPI
React
TypeScript
PostgreSQL
Rust