LangGraph: wire the agent
A puzzle game. Connect the arrows between an agent’s steps, run the tests, and learn how LangGraph agents loop, stop and wait for people.
How it works
- 1LangGraph builds an AI app as a graph: steps (nodes) that share one state, joined by arrows (edges). Some arrows depend on what just happened.
- 2Each level leaves some arrows open. Pick where they go, then run the tests — you watch every test walk through your graph.
- 3Four levels, four patterns: a plain graph, a tool loop, a retry limit, and pausing for a person’s approval.
- 4No AI model is called. The steps are tiny bits of code, so the game is about the graph itself — which is exactly what LangGraph adds.
More demos
All demosQueryLite: a SQL database
A SQL database written from scratch — parser, B+ tree indexes, a query planner, joins and transactions — running live on 22,000 rows in your browser.
Raft consensus, live
The algorithm that keeps etcd and Kubernetes consistent. Five servers elect a leader and replicate a log — crash them and split the network while it runs.
Live collaborative editor
Real-time editing with no server. Three devices share a note — take one offline, edit everywhere, reconnect, and they merge. Open a second tab and it syncs live.