Documentation

How Handoff works and how to get the most out of it. Short pages, written for the engineer who will own the suite.

From an empty organisation to a first supervised run in about ten minutes.

How agent and human share one browser without racing each other.

How manual actions during a takeover become reviewable, re-runnable steps.

Why replays are fast, model-free and resistant to the usual sources of flakiness.

Binding Anthropic, Google, OpenAI-compatible endpoints and what gets probed.

Run Ollama, vLLM or LM Studio behind your firewall without opening a port.

What the exported spec contains and how to run it in your own CI.

The paper trail: how the RTM and cycle report are derived and what an incident is.

What the stack needs on a plain Linux server.