> ## Documentation Index
> Fetch the complete documentation index at: https://docs.repli5.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Projects & Data Library

> Upload datasets, organize team projects, manage datapoint tags, and export filtered training archives on the Repli5 platform.

## Working with Projects

On **the Repli5 platform**, **Projects** let you group datasets, saved workflow templates, and augmentation runs into dedicated workspaces for a specific product line, part family, or team—while still allowing you to work standalone under **All Projects** whenever you prefer.

### Creating & Switching Projects

* **Create a Project:** In the left sidebar under **Projects**, click **+ New**, enter a project **Name** and optional **Description**, and click **Create project**.
* **Switch Projects:** Click any project in the sidebar (or click **More projects** / use the search box if you have more than 5 projects).
* **Automatic Scoping & Last-Used Workflow:** Selecting a project scopes your **Data Library**, saved **Templates**, and **Augment History** to that project, and automatically opens the workflow that was last used in that project. Click **All Projects** in the sidebar at any time to return to your unscoped personal view.

### Managing Project Datasets, Team Members & Settings

Click the **`⋮`** action menu next to any project in the left sidebar to:

* **Manage Datasets:** Attach existing datasets from your Data Library to the project or remove datasets from the project.
* **Manage Users:** Invite teammates by email and assign roles (*Owner* or *Member*) so your team can share datasets, templates, and augmentation runs.
* **Rename / Edit Project:** Update the project's name or description.
* **Pin to Top:** Keep frequently used projects pinned at the top of your sidebar list.
* **Duplicate Project:** Clone a project's setup into a new project.
* **Delete Project:** Remove the project workspace when it is no longer needed.

***

## Uploading Datasets to the Data Library

Navigate to **Data Library** in the left sidebar to upload and manage datasets on **the Repli5 platform**.

### Supported `.zip` Upload Formats (Up to 10 GB)

You can upload a `.zip` archive in either **Images-Only** format (for **Defect Generation** or **Variation Generator**) or **Multi-Modal** format:

| Folder in `.zip` | Requirement | Supported Files | Description |
| :- | :-: | :-: | :- |
| **`rgb/`** or **`rgb_raw/`** | **Required** | `.png`, `.jpg`, `.jpeg` | Primary RGB source images (e.g., defect-free `OK` part photos or close-up object photos). |
| **`mask_raw/`** | Optional | `.png` | Raw segmentation masks matched by filename stem. |
| **`instances/`** | Optional | `.png` | Instance segmentation masks matched by filename stem. |
| **`semantics_rgb/`** | Optional | `.png` | Colorized semantic segmentation maps. |
| **`depth/`** or **`depth_carla/`** | Optional | `.png` | Metric depth maps. |
| **`bbox_2d/`**, **`bbox_3d/`**, **`keypoints/`** | Optional | `.txt` | 2D boxes, 3D boxes, or pose keypoints matched by filename stem. |

***

## Inspecting, Tagging & Exporting Datasets

Opening any dataset in the **Data Library** gives you full control over individual datapoints before and after augmentation:

1. **Full-Resolution Inspection & Overlays:** Click any datapoint to inspect `rgb_raw` vs. `rgb_out`, toggle segmentation masks or bounding boxes, and zoom/pan at full resolution.
2. **Datapoint Tagging & Filtering:**
   * Create custom tags and assign them to individual datapoints (or mark unwanted samples with the `reject` tag).
   * Use the **Tag Filter Bar** at the top of the dataset view to include or exclude specific tags.
3. **Exporting & Downloading Datasets:**
   * Click **Download** (or **Export**) on any dataset in the Data Library.
   * When exporting a tagged dataset, you can include or exclude specific tags (any datapoints tagged `reject` are excluded by default) and download a ready-to-train `.zip` archive containing your images, segmentation masks, `annotations_coco.json`, and `tags.json`.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.