Create

Create a project in which to store images and train a model.

You can create a project using the Roboflow Python SDK.

Python SDK

To create a project in Roboflow, use the rf.create_project() method

rf.create_project(
    project_name="project name",
    project_type="project-type",
    project_license="project-license",
    annotation="annotation-group"
)

REST API

The POST /:workspace/projects endpoint allows you to create a project in a workspace.

Example Request

$ curl --location 'https://api.roboflow.com/roboflow/projects?api_key=$ROBOFLOW_API_KEY \
--header 'Content-Type: application/json' \
--data '{
    "name": "Sharks Dataset",
    "type": "object-detection",
    "annotation": "sharks"
}'

Parameters

- name (string) - [Required] the name of the project
- type (string) - [Required] the project type. Accepted values:
    - object-detection
    - single-label-classification
    - multi-label-classification
    - instance-segmentation
    - semenatic-segmentation
- license (string) - Required for public workspaces unless the value is "Public". Accepted values:
    - Public Domain
    - MIT
    - CC BY 4.0
    - BY-NC-SA 4.0
    - OBdL v1.0
    - Private
- annotation (string, alphanumeric) - [Required] the name of the annotation group
- group (string) - Id of the group to add this project to 

Example Response

{
    "id": "roboflow/sharks-dataset",
    "type": "object-detection",
    "name": "Sharks Dataset",
    "created": 1688739471567,
    "updated": 1688739471567,
    "images": 0,
    "unannotated": 0,
    "annotation": "sharks",
    "versions": 0,
    "public": false,
    "splits": {},
    "colors": {},
    "classes": {},
    "icon": null
}