> ## Documentation Index > Fetch the complete documentation index at: https://docs.orbitagents.xyz/llms.txt > Use this file to discover all available pages before exploring further. # How it all fits together > The full lifecycle, in one loop. Everything in Orbit serves one loop. Here's how the pieces connect — and why each run makes the next one smarter. ## The loop ``` You file a decision → Agent runs → Reads decision via orbit_heartbeat → Does the work → Files output via orbit_result → Output lands in the Vault → Appears in your Inbox as new → You review and approve → Stays in the Vault permanently → Future agents read it as knowledge ``` ## Why this compounds Run that loop across multiple agents, across multiple tools, every day — and your operation stops resetting. Each output becomes context for the next agent. Each decision governs every future run. Each approved item makes your Vault smarter. A traditional AI stack is a set of islands that each start from zero. Orbit turns it into a single system with shared memory and shared direction. ## The pieces in the loop * **[Decisions](/concepts/decisions)** set direction every agent reads. * **[Agents](/concepts/agents)** do the work and file results. * **The [Vault](/concepts/vault)** stores everything permanently. * **The [Inbox](/concepts/inbox)** is where you review new work. * **[Knowledge](/concepts/knowledge)** is what graduates to ground truth. * **The [Briefing](/concepts/briefing)** is your daily view of it all. Connect your first agent and watch it run.