> ## Documentation Index
> Fetch the complete documentation index at: https://docs.athina.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Configure an Annotation Project

> Learn how to set up and configure an annotation project in Athina AI.

Athina AI enables you to launch fully managed annotation projects to collect high-quality labeled data. This guide walks through the full setup process—from project creation to annotator assignment.

***

## Step 1: Create a Project

Navigate to the **Annotations** page and click the **+ Create project** button at the top right.

![Create Annotation Project](https://mintlify.s3.us-west-1.amazonaws.com/athinaai/images/annotations/configure_project/create_project_button.png)

On the creation page, fill in the required project information:

* **Name** – Internal name of your annotation project.
* **Description** – Optional. Helps team members understand the scope.
* **Instructions** – Optional guidelines shown to annotators for consistent labeling.

![Project Creation Form](https://mintlify.s3.us-west-1.amazonaws.com/athinaai/images/annotations/configure_project/create_annotation_form.png)

***

## Step 2: Select Dataset

Choose a dataset from the dropdown.\
Currently, only existing datasets can be selected.

* Supported file formats for datasets: `.jsonl`, `.json`, and `.csv`
* If needed, upload the dataset from the [Datasets](/datasets/overview) page beforehand.

***

## Step 3: Define Annotation View

Annotation View determines what annotators see and how they interact with the data.

* You can **select a template** (predefined configuration) or **create a new view**.
* Define:
  * **Which fields** are visible (e.g., input, output, metadata).
  * Whether fields are **editable**, **required**, or **markdown-enabled**.
  * **Labels** to be collected (e.g., category, number, comment).

> View configurations can be reused across multiple projects.

***

## Step 4: Set Completion Criteria

Define how many annotators must label each datapoint.

* **Minimum required annotations per datapoint**:\
  Determines how many independent annotations are needed before a datapoint is considered complete.
* Toggle: **Skip datapoint if minimum requirement is met** – allows annotators to skip fully-labeled entries.

<Info>
  If you set this value to `2`, each datapoint must be labeled by 2 different annotators to be marked as complete.
</Info>

***

## Step 5: Choose Assignment Method

Select how to distribute annotation workload:

* **Entire dataset** – All annotators label the full dataset.
* **Split into subsets** – Automatically or manually divide the dataset and assign different subsets to annotators.

<Tip>
  Use subsets if you want to avoid duplicate annotation or manage workloads among many contributors.
</Tip>

***

## Step 6: Assign Annotators

* Search and add your team members as annotators.
* Depending on the assignment method, they will receive:
  * All datapoints (entire dataset) or
  * Only their subset (split mode)

***

## Step 7: Launch

When you're done configuring, click **Launch project**. Annotators will be notified and can begin labeling immediately.

***