Core Concepts Overview

Understand PuppyOne's data model and core architecture.

Data hierarchy

PuppyOne organizes your data in a three-layer structure:

Organization
└── Project
    └── Content Nodes (Mut Merkle tree)
        ├── Folder
        │   ├── JSON file
        │   ├── Markdown file
        │   └── Uploaded file
        └── ...

Organization

An Organization is the top-level management unit, typically representing your team or company.

• Manage members and roles
• Contain multiple Projects
• Centralize billing and permission management

Project

A Project is the namespace for a specific set of business data.

• For example: "Product Knowledge Base", "Support FAQ", or "Engineering Docs"
• Each Project has its own independent content tree, stored in a Mut Merkle tree backed by S3
• You can configure project-level access control and membership

Content Node

A Content Node is PuppyOne's core data structure: a tree made up of folders, JSON files, Markdown files, and uploaded files, all identified by path.

Unlike a traditional table model, the content tree supports arbitrary nesting and can hold both structured data and unstructured documents, much like a cloud file system.

Core concepts

Access Points: the single entry point for integrations

PuppyOne uses the concept of an access point to manage every external integration in one place:

Access Point
├── Sync Source   - Pull data from Notion, GitHub, Gmail, and more
├── Agent         - AI agent with model and tool bindings
├── MCP Endpoint  - Expose data through MCP
├── Sandbox       - Isolated code execution environment
└── Filesystem    - Bidirectional local folder sync

Whether you are connecting a Notion source or creating an MCP endpoint for Cursor, you use the same interface:

puppyone access add notion <url>         # Add a Notion data source
puppyone access add mcp "My MCP"         # Create an MCP endpoint
puppyone access add filesystem ~/docs    # Mount a local folder
puppyone access ls                       # List all access points

Learn more