> ## Documentation Index
> Fetch the complete documentation index at: https://docs.platform.nora.my/llms.txt
> Use this file to discover all available pages before exploring further.
# Core concepts
> Workspaces, Flows, and Blocks — the vocabulary you'll use everywhere in Nora.
Nora has a small vocabulary. Learn these four terms and the rest of the product reads itself.
## Workspace
The top-level container. A **Workspace** holds your Flows, data sources, teammates, credits, and settings. One account can belong to several workspaces (personal, team, client).
Everything in Nora — a Flow, a Signal, a Trace — lives inside exactly one Workspace. Workspaces are isolated from each other.
Workspaces in depth
## Flow
A **Flow** is one agent workflow — a canvas of Blocks connected together. Think of it as one product, one job.
You'll typically have one Flow per use case: a support assistant, an internal Q\&A bot, a research helper. Flows are versioned; you can publish, roll back, and hold changes for approval.
Flows in depth
## Block
Blocks are what you drop on a Flow canvas. There are five kinds:
| Block | What it does |
| -------------- | --------------------------------------------------------------------------------------- |
| **Agent** | The reasoning core — a model with a role, instructions, and tools |
| **Tool** | An action the agent can call — API request, function, code snippet |
| **Action** | A follow-up step that runs after the agent — send email, write to a sheet, notify Slack |
| **Trigger** | What starts the Flow — chat message, webhook, schedule |
| **DataSource** | Documents, spreadsheets, or a database the agent can search |
Blocks connect to each other on the canvas. The Agent decides which Tools to call; Actions and Triggers wrap the Flow's inputs and outputs.
Building with Blocks
## Memory
Your agents remember. **Memory** is where useful facts and patterns from past runs are kept and re-used — so an agent doesn't relearn the same things every session.
You control what's remembered and can inspect it at any time.
Memory in depth
***
## How they fit together
```
Workspace
└── Flow "Support assistant"
├── Trigger (chat message from your app)
├── Agent (support-agent, GPT-4-class model)
│ ├── Tool (look-up-account)
│ ├── Tool (issue-refund)
│ └── DataSource (product docs)
└── Action (log to CRM)
```
Every answer produced by that Flow is a **Trace**. If a user thumbs-down the answer, or if a rule flags it, it becomes a **Signal** in your Improve view.