---
title: "System Initialization"
date: 0001-01-01
summary: "System Initialization #  Initialization API #  Initialize the Coco Server with an admin account and optional LLM configuration. This endpoint can only be called once — subsequent calls will return an error.
Parameters #     Field Type Required Description     name string Yes Server name (becomes “{name}’s Coco Server”).   email string Yes Admin user email address.   password string Yes Admin user password."
---


# System Initialization

## Initialization API

Initialize the Coco Server with an admin account and optional LLM configuration. This endpoint can only be called once — subsequent calls will return an error.

### Parameters

| **Field**          | **Type**  | **Required** | **Description**                                                  |
|--------------------|-----------|--------------|------------------------------------------------------------------|
| `name`             | `string`  | Yes          | Server name (becomes "{name}'s Coco Server").                    |
| `email`            | `string`  | Yes          | Admin user email address.                                        |
| `password`         | `string`  | Yes          | Admin user password.                                             |
| `llm.type`         | `string`  | No           | LLM provider type, e.g., `openai`, `ollama`.                    |
| `llm.endpoint`     | `string`  | No           | LLM API endpoint URL.                                            |
| `llm.default_model`| `string`  | No           | Default model name, e.g., `deepseek-r1`.                        |
| `llm.token`        | `string`  | No           | API token for the LLM provider.                                  |
| `llm.reasoning`    | `boolean` | No           | Enable reasoning mode for the model.                             |
| `language`         | `string`  | No           | Language setting: `en-US` or `zh-CN` (default: `zh-CN`).        |

```shell
//request
curl -H 'Content-Type: application/json' -XPOST http://localhost:9000/setup/_initialize -d'{
  "name": "Coco",
  "email": "hello@coco.rs",
  "password": "mypassword",
  "llm": {
    "type": "ollama",
    "endpoint": "http://localhost:11434",
    "default_model": "deepseek-r1"
  },
  "language": "en-US"
}'

//response
{
  "status": "ok"
}
```

## Initialization UI Management

When entering the coco server, it will check whether it has been initialized. If not, it will enter the initialization page.

### Create a user account

Set up a new user account to manage access and permissions.

{{% load-img "/img/initialization/step-1.png" "initialization step 1" %}}

### Connect to a Large Model

After integrating a large model, you will unlock the AI chat feature, providing intelligent search and an efficient work assistant.

{{% load-img "/img/initialization/step-2.png" "initialization step 2" %}}

You can also click `Set Up Later` to skip and configure it later after entering the system.