# AgentCube documentation: connect AI to Oracle EPM The starting point for installing, configuring, and operating AgentCube — where to go first and how the rest of these docs are laid out. Welcome to the AgentCube documentation — the **owner's manual** for the team that installs, configures, and runs AgentCube. ## Start here 1. **[Introduction — What is AgentCube?](/docs/introduction/)** — what it is, who it's for, and why it works the way it does. Start here if AgentCube is new to you. 2. **[How it works](/docs/how-it-works/)** — the Context-Engine-as-hub story, following a single question from the user through to Oracle and back. 3. **[Quickstart](/docs/quickstart/)** — the shortest path from zero to a working query: deploy, register, connect your AI, and ask your first question. 4. **[Installation](/docs/installation/)** — once you know the shape of it, deploy the Context Engine and the connectors. ## The rest of the manual - **[Architecture](/docs/architecture/)** — system topology, the Model Context Protocol, connectors, trust boundaries, and security posture, for the architect evaluating AgentCube before install. - **Connect your AI** — point your AI platform at the Context Engine: [Claude](/docs/installation/connecting-claude/), [Copilot](/docs/installation/connecting-copilot/), [ChatGPT](/docs/installation/connecting-chatgpt/), or [Gemini](/docs/installation/connecting-gemini/). - **[Administration](/docs/administration/)** — day-to-day operation: adding connectors and data sources, health monitoring, scheduled jobs, users, auth modes, certificate rotation, upgrades, rollback, and troubleshooting. - **[Security](/docs/security/)** — the full security posture: source-native authorization, per-user isolation, data handling, and the shared-responsibility model. - **[FAQ](/docs/faq/)** — common questions about security, compatibility, and authentication.