> For the complete documentation index, see [llms.txt](https://docs.biodock.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.biodock.ai/start/path-2-train-a-fully-automated-ai-model/set-up-your-project.md).

# Set up your project

*What you'll get: a project with your images in it, a segmentation type, a tile size and your first class, ready to label. About 5 minutes.*

## 1. Create a project

<figure><img src="https://3806122466-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk3DfNQKoMy1JXdaDLvpG%2Fuploads%2Fgit-blob-ec6bc6f59afbfa924d588ccbdaca05dad0a35f28%2Fdocs-start-create-project.gif?alt=media" alt="On the Projects page, Create a new project is clicked, the project is named Nuclei counter and Create project is clicked; on the new project, Add images to begin opens the file picker, the sample data folder is ticked and Add 1 file to project is clicked"><figcaption><p><strong>Create a new project</strong> → <strong>Create project</strong> → <strong>Add images to begin</strong> → tick a folder → <strong>Add 1 file to project</strong>.</p></figcaption></figure>

1. Click **Projects** in the sidebar, then the first tile, **Create a new project**.
2. Under **Project name**, type a name you'll recognise, such as `Nuclei counter`. You can rename it later.
3. In a [workspace](https://docs.biodock.ai/workspaces/workspaces/sharing-and-visibility.md), choose **Who can see it**. **Everyone in** *your workspace* makes it an open project, where every member can label. **Only people with access** keeps it to you until you share it.
4. **Color** and **Pattern** only set the project's thumbnail. Click **Create project**.

<figure><img src="https://3806122466-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk3DfNQKoMy1JXdaDLvpG%2Fuploads%2Fgit-blob-361283e6a8f4490baa7a37021b19d623b34384be%2Fv2-projects-page.png?alt=media" alt="The Projects page with the Create a new project tile"><figcaption><p>The <strong>Projects</strong> page. The first tile is <strong>Create a new project</strong>.</p></figcaption></figure>

## 2. Add your images

A project takes its images from **Files**, so [upload them there](https://docs.biodock.ai/files/uploading-images.md) first. Upload them into the place you will work in: files never move between your personal account and a workspace.

1. On the project page, click **Add images to begin**. Later, use **Add from files** on the **Data** tab.
2. Tick the files or folders you want. Ticking a folder takes everything inside it, sub-folders included. Click a folder's name to open it.
3. Click the add button at the bottom, for example **Add 1 file to project** (a folder counts as one).

The **Data** tab now lists your images. The project gets its own copy of each one: the originals stay in Files.

{% hint style="info" %}
**Every image in a project needs the same number of channels.** Train on images taken at the same zoom you will analyse later, with a mix of bright, dim, crowded and sparse.
{% endhint %}

## 3. Check your channels

Images with three channels or fewer need no setup: Biodock shows them together and the model learns from all of them.

With more than three channels, or if a channel is too dim, set up the channels first: on the project page, click **Settings** → **Appearance & channels**. See [Project settings](https://docs.biodock.ai/projects-and-models/project-settings.md).

## 4. Tell Biodock what to look for

<figure><img src="https://3806122466-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fk3DfNQKoMy1JXdaDLvpG%2Fuploads%2Fgit-blob-f7bb5d34e779af4504a2701267e764b68646efec%2Fdocs-start-first-size-group.gif?alt=media" alt="Label data is clicked, Detect each object (instance) is selected and Next is clicked, a tight box is dragged around one nucleus, the recommended tile size appears and Confirm is clicked, then the class is named Nucleus in Add your first class and Done is clicked"><figcaption><p><strong>Label data</strong> → <strong>Detect each object (instance)</strong> → box one object → <strong>Confirm</strong> → <strong>Add your first class</strong> → <strong>Done</strong>.</p></figcaption></figure>

On the project page, click **Label data**. The first time, Biodock asks three questions.

#### The segmentation type

In **Select a segmentation type**, choose one and click **Next**:

* **Detect each object (instance)** for things you count and measure one by one: cells, nuclei, organoids. It is marked **Most common**.
* **Detect each region (semantic)** for areas, where separate objects don't matter: inflamed tissue, or how much of the dish the cells cover.

#### The tile size

Biodock cuts every image into square tiles, and the model looks at one tile at a time. You label tile by tile too.

* **For objects:** zoom in, find an average-sized example (not the biggest), and drag a tight box around it. Biodock shows the **Recommended** size and sets the **Size** dropdown to it. **Clear selection** lets you box another one.
* **For regions:** hover the yellow circle in the centre of each tile to compare it with an average-sized region, and pick a **Size**.

Hold **Space** and drag to move around. If the grid is hard to see, change its **Preview** colour. Then click **Confirm**.

It is **very important** to get the tile size right: it lets the model see your objects at the right scale. You can change it later, but tiles you have already finished may need checking again.

#### Your first class

In **Add your first class**, type a **Class name**, such as `Nucleus`, and pick a **Class color** you can see on your images. Click **Done**.

## The labeling screen

The labeler opens. On the left are the image picker, **Channels**, your size group with its classes, and **Labeled objects**. In the middle is your image, cut into tiles.

A size group is a tile size plus the classes that share it. Its type can't change later; to find regions as well as objects, add a second size group.

Continue to [Label and train](https://docs.biodock.ai/start/path-2-train-a-fully-automated-ai-model/label-and-train.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.biodock.ai/start/path-2-train-a-fully-automated-ai-model/set-up-your-project.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
