English
Data Management
Overview

Manage Data

Create, edit, and organize your data in PuppyOne.


Why data management matters

Imported raw data often needs:

  • Cleanup - Remove irrelevant fields and fix errors
  • Reorganization - Reshape JSON so Agents can query it more easily
  • Enrichment - Add metadata and tags

PuppyOne data model

In PuppyOne, data is organized as a Content Node tree, similar to a file system:

Project
├── 📁 products/              ← Folder node
│   ├── 📄 catalog.json       ← JSON node
│   └── 📄 intro.md           ← Markdown node
├── 📁 docs/
│   ├── 📄 readme.md
│   └── 📁 api/
│       └── 📄 endpoints.json
└── 📎 logo.png               ← File node
Node typeDescription
FolderA folder used to organize hierarchy
JSONStructured data with objects, arrays, and nesting
MarkdownRich text content for docs and knowledge bases
FileBinary files such as images and PDFs

Each Content Node has its own path, access control, and version history.

For more details, see Core Concepts - Content Node.


Core features

JSON tree editor

Visually edit JSON data:

  • Navigate nested structures in a tree
  • Click any value to edit it
  • Add new fields or array items
  • Delete nodes you no longer need

Detailed guide →

Version control

Every change is automatically saved as a version:

  • View change history
  • Compare different versions (Diff)
  • Roll back to any version in one click
  • Create folder-level snapshots

Learn about version control →


Data structure recommendations

To help Agents use your data more effectively, we recommend:

Recommended structure

{
  "products": [
    {
      "id": "SKU-001",
      "name": "Widget Pro",
      "price": 99.99,
      "category": "Electronics"
    }
  ],
  "categories": ["Electronics", "Accessories", "Services"]
}

Characteristics: flat, clear field names, consistent types

Not recommended

{
  "data": {
    "items": {
      "group1": {
        "subgroup": {
          "products": [...]
        }
      }
    }
  }
}

Problem: the nesting is too deep, so Agents have a harder time locating data accurately

More best practices →


Next steps