← nekolabs

Artifacts

Things we made along the way. Self-contained, interactive, one page each.

Wiring Claude Hooks into an app

How Claude Code hooks work and how to build on them: where they fire in the agent loop, the JSON-in exit-code-out contract, an observer script that never blocks, a guard that does, and the wiring that turns a fleet of sessions into one event stream with a back-channel.

claude-code hooks explainer

Tokens — How AI Reads Your Words

An AI literacy module. Type a sentence and watch a teaching tokenizer break it apart, learn the four-characters-per-token rule of thumb, then dial a job through a cost meter across three Claude models.

ai-literacy tokens interactive