Documentation

Your team and your AI agents work from the same context: tasks, their whole history and the time they took live in your repo as an append-only journal — one file per event, right next to the code.

npm install -g kadence
kadence init

In a terminal, init walks you through the setup; run by an agent, it asks nothing. No server. No account. No network.

What it is for

Your code says what exists. git log says when it changed. Neither says what was tried and abandoned, why a task is blocked, or what the team agreed on Tuesday.

That gap costs a person a few minutes. It costs an AI agent the whole session: every new one starts from scratch, re-reads the same files, and asks the question you answered yesterday.

kadence keeps that missing layer as events, committed with the code. One call returns the whole state of a piece of work — status, blockers, comments, hours, and every step that led there.

Where to go next

• Quickstart — the first ten minutes, the teammate after them, and how to leave.
• How it works — events, folding, and why state is derived.
• CLI reference — every command.
• Recording why — decisions as events, and documents linked to work.
• For AI agents — --json, the schema, and the guide init writes.
• Proof — the merge claim, and what an answer costs an AI agent.
• Decisions — every ADR, with what was measured.

What it costs you

These are tests. They fail the build on regression, which is why they are still true. The last row is Probe C, which measured what an answer costs when the journal behind it has grown to half a megabyte.