Concepts
Before clicking anything, it helps to know the handful of ideas everything in Studio is built from. Most of the UI is just these terms wearing different hats depending on which screen you're looking at.
Core concepts, in the order you'll meet them
| Term | What it actually is |
|---|---|
| Collection | The top-level container. One isolated graph — its own files, schema, links, snapshots. Roughly "a database" or "a workspace." |
File (.kgl) |
A collection is split into one or more .kgl files. Each file holds items of (usually) one type — it's the closest thing to "a table." Folders are just / in a filename (e.g. sales/deals.kgl) — there's no separate folder object. |
| Schema / Type | A named shape (e.g. Company) with a list of typed fields. Defined once, applied to every item of that type. |
| Field | One typed column on a schema — text, number, select, image, relation, etc. See the full list in the field type reference. |
| Item / Node / Row | One record. Same thing, three names depending on context (the grid calls it a row, the graph calls it a node, the format calls it a node). Has a name (unique within its file), one or more types, field values, an optional rich-text body, tags, and links. |
| Link / Relationship | A directed, typed edge from one item to another — Contact --WORKS_FOR--> Company. Links can cross files and even cross collections. Can carry key/value properties (e.g. since: 2023). |
| Relationship type (RelType) | The label on a link (WORKS_FOR, FOR_COMPANY, …), optionally declared in the schema with a From/To type constraint. Ad-hoc rel types are also allowed. |
| Tag (snapshot) | Not a free-form label on an item — a point-in-time copy of the entire collection (every file). Used as a manual backup/checkpoint you can view read-only or restore. If you want per-item labels, that's the tags: list on a node — a different thing entirely. |
| Enrichment (queued fields) | For image/pdf fields with "automated fields" turned on: an AI worker fills in a Description and Tags (as graph links to a Classification type) after upload. Runs asynchronously — the cell is blank until the worker finishes, then Studio refreshes automatically. |
Starting a collection
- From the Collections page (or the + New collection button), choose a starting point:
- Blank — nothing pre-built.
- Templates — nine bundled starters (People & Organizations, Project Tracking, CRM & Sales Pipeline, Software Bug Tracker, Content & Blog CMS, Product Inventory & Suppliers, Event Planning, Knowledge Base & Research, Recruiting & Hiring). Each ships with a working schema and a few example rows already linked together — good for seeing a finished shape before building your own.
- Clone an existing collection — copies its schema only (or schema + data via the collection card's "Clone schema" action), not its data.
- Give it a name; the URL slug auto-fills from the name (editable).
- Create. You land in the editor with that collection active.
The worked example at the end of this guide builds
from Blank, on purpose — it's the same shape as the bundled CRM & Sales Pipeline
template, but walking through it yourself is how the next two sections actually stick.