Skip to main content
A gavAI app is five ideas snapped together: pages assembled from composable blocks, capabilities that wire those blocks to real-world actions, a schema the runtime validates writes against, and a multi-tenant runtime that keeps every workspace isolated. By the end of this section you will know what each piece is, why it exists, and which page to read when you need to go deeper. For who the platform is for and what is included at each entry point (Builder, MCP, code), see Who this is for. The concepts below describe how the platform works once you’ve chosen an entry point. The pieces aren’t independent — they compose. A page is a JSON document; the document declares a schema; blocks in the document bind to schema tables and invoke capabilities; the runtime renders the document and dispatches the capability calls on per-tenant infrastructure. You can read these pages in any order, but the order below is the one that builds the picture fastest.

​
A suggested reading order

If you’re new to gavAI, read them top-to-bottom: pages first (so the artifact is concrete), then capabilities (what pages can do), then schema (what they read and write), then the multi-tenant model and architecture (where it all runs). The Builder page is independent — read it whenever you want to see the visual surface a non-technical owner uses. If you’re integrating from the outside, start with Architecture and Multi-tenant model, then come back for the page model when you need to read or patch a document.