> For the complete documentation index, see [llms.txt](https://cleyrop.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://cleyrop.gitbook.io/docs/v4.7-en/getting-started/tutoriel-dintroduction.md).

# Introduction tutorial

This tutorial guides you step by step through discovering Cleyrop's main features.

This tutorial guides you step by step through the main features of Cleyrop.

In a few minutes, you will:

1. Log in to the platform,
2. Upload a dataset,
3. Create a transformation,
4. Use Le Druide,
5. Build a first Agent that queries your data or your documents.

## Login & authentication

### First login <a href="#premiere-connexion" id="premiere-connexion"></a>

You received a first-login email to set up your account:

* Choose an **password** to securely access your account.

{% hint style="info" %}
The subject line is "Welcome to Cleyrop"
{% endhint %}

Later, to log in, you need:

* your Cleyrop instance URL
* username
* password

#### Expired link <a href="#premiere-connexion" id="premiere-connexion"></a>

The link expires after 12 hours; if it has expired, ask your platform manager to resend you a login email.

### MFA setup

If MFA (*multi-factor authentication)* is enabled for your instance and your account, you must first **set up access from your device** (mobile or computer).

1. Scan the QR code from your device using one of the dedicated apps on your mobile: Google Authenticator, Microsoft Authenticator, Free OTP
2. Name your device
3. Then set up your password

<figure><img src="https://4159280578-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FMHNFH6aA4j5ucH4XFBXy%2Fuploads%2Fgit-blob-1e31ec2791fc78b7a2df2edc8e0cff49035ac174%2Fimage.png?alt=media" alt="" width="375"><figcaption></figcaption></figure>

## Tutorial 1 - Upload a dataset

**Objective:** Import a data file (CSV, Excel, JSON, etc.) to make it usable on the platform.

**Steps**

1. From the home page of a project or from **My Studio**, go to **Library > Datasets** and click on **Create**.
2. Choose **Local import**.
3. Select or drag your file and follow the creation steps: reading settings, [schema policy](/docs/v4.7-en/data-and-ai-project/datasets/politique-de-schema.md), metadata.
4. Once the import is complete, the dataset appears in the project's dataset list.
5. View the preview from the dataset page to check consistency.

**Good to know**

* Imported datasets are stored in a secure project area.
* You can define a **sensitivity** level (Internal, Restricted, Sensitive).
* To import multiple files and get one dataset per file, choose **Create > Multiple datasets**.

## Tutorial 2 - Create a transformation

**Objective** : Clean, filter, or enrich your data to prepare it for analysis.

**Steps:**

1. Open the Dataflows tab of your project or My Studio.
2. Click Create a Dataflow then add your dataset as input.
3. Choose a transformation:
   * Low code to filter, join, or aggregate columns,
   * Python / PySpark / SQL to write your own scripts.
4. Click Run to launch the transformation.
5. Once the job is finished, view the output dataset.

**Good to know**

* Execution sessions are parallelized on your cluster (Spark or Python).
* You can view execution logs to diagnose possible errors.

## Tutorial 3 - Use Le Druide

**Objective** : Discover how to use Le Druide, Cleyrop's secure conversational AI assistant, to interact with your data, your documents, and your tools.

**Steps**

1. Click on **New Chat** from the side menu.
2. Add **an attachment** (PDF, Word, Excel, etc.) to analyze a document in the conversation.
3. Enter your question (e.g. “Summarize this document in 5 points.”) in the chat field to get an instant answer.
4. Activate a **Tool** in the dedicated panel, such as **Cleyrop Internet** for online search or **Cleyrop Dataset** to query a dataset in natural language.
5. Ask for a **chart** or a **document** (Word, PDF, PowerPoint) from your exchanges, then download it.
6. Consult the history or start a new session to begin again from a blank context.

**Good to know**

* Fully sovereign environment and models.
* The full range of Le Druide's capabilities is described in [Secure Chat](/docs/v4.7-en/le-druide/chat-securise.md).

## Tutorial 4 - Create an Agent

**Objective** : Creating an Agent means configuring Le Druide for a specific business need, so that it queries your datasets or your documents.

**Option 1 — Agent on your datasets**

1. From the **AI Studio** of your project, tab **Agent**, click **Create an Agent**.
2. Give it a name and write its **instructions** (the expected role, tone, response rules).
3. Select the Tool **Cleyrop Dataset**, then choose one or more project datasets within its knowledge scope.
4. Save, then open the Agent in Le Druide and ask your first questions (“What is the number of rows?”, “Which customers have the most orders?”).
5. The Agent translates your questions into SQL, executes the query and returns the results, which you can download or display as a chart.

**Option 2 — Agent on your documents**

1. Drop your documents (PDF, DOCX, TXT, etc.) into the **Work Data** of the project (Library > Work Data).
2. From **Library > Corpus**, create a corpus from these files and wait for it to reach the *Valid*.
3. Create an Agent with the Tool **Cleyrop Corpus** and select this corpus within its knowledge scope.
4. Ask your questions: “What does the report say about 2024 performance?”, “What risks are identified?”. The Agent answers by citing its sources.

**Good to know**

* An Agent can combine datasets, corpora, reference files and Tools created by your colleagues in The Factory.
* An Agent saved without publication remains usable by project members with the status **Test** ; publish it once validated. The details are described in [Agents](/docs/v4.7-en/data-and-ai-project/studio-ai/agents.md).