# How it works

{% hint style="success" %}
This is the official documentation of the `@forestadmin/agent` Node.js agent.
{% endhint %}

Before you start writing a single line of code, it’s a good idea to get an overview of how Forest Admin works. The magic lies in its architecture.

Forest Admin provides you with:

* An API, hosted on your server, to retrieve your data. We call it the **Agent**. It can be a **standalone project** or embedded **within your app**.
* A user interface, to access and manage your data from your browser. This **Forest Admin User Interface** is built and managed through resources hosted on Forest Admin's servers.

![The agent is a Node.JS REST API hosted on your servers](https://3861847666-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F9UN5oBJhgzLadOqi7jx6%2Fuploads%2Fgit-blob-e9d088f915c6815710d294bc53dfbb80443c6957%2Fhow-it-work-architecture.png?alt=media)

{% hint style="info" %}
For a more in-depth explanation of Forest Admin's architecture (the Node.JS agent version), please read the [following article](https://medium.com/forest-admin/a-deep-dive-into-forest-admins-architecture-and-its-benefits-for-the-developers-who-trust-it-1d49212fb4b).
{% endhint %}

### The agent

The agent is generated upon installation and **hosted on your end**.

It includes an API allowing you to **translate calls made from the Forest Admin UI into queries** to your database (covering actions such as CRUD, search & filters, pagination, sorting, etc.).

It also provides the Forest Admin servers with the information needed to build the User Interface (the **Forest Admin Schema**). This information includes table names, column names and types, and relationships. It is sent when you run your agent within a file called `forestadmin-schema.json`.