STARNET DOCS
// everything you need to run your station
StarNet is a local-first desktop harness: you create AI agents, organize them into a living pixel-art station, and the station's layout is the workflow they run. These docs cover installing it, connecting a model provider, and using every major system.
Install on Windows, macOS, or Linux, get past the first-run OS warnings, and run your first agent.
Bring your own OpenRouter key, sign in with OAuth, and store service keys your agents can use.
Rooms, hallways, and placed objects — how the layout you draw becomes real capability boundaries.
Recruiting, agent classes, the dossier, and running many agents concurrently.
Let the station keep working while you're away — on a leash you define, with honest reporting.
The how, the what, and the when of repeatable agent work.
MCP connectors, messaging channels, schedules, and using StarNet alongside other harnesses.
Concept map for migrating users — and how to keep your existing harness in the loop.
Install warnings, updater behavior, Linux libraries, uninstalling, and where to get help.