Create a new project

To create a new project programmatically, you can use the DVC Studio create-project API. Authentication is required using a token, which can be generated from the user's profile settings in DVC Studio.

Endpoint: api/create-project
HTTP Method: POST
Content-Type: application/json

Authentication

The request should contain following header containing DVC Studio client access token with PROJECT scope for authorization.

Parameters

Here are the possible values for each parameter in the request json:

• repo_url (required) - the URL of the repository to create the project from. This field is mandatory.
• public (boolean, default: false) - specifies whether the project should be public (true) or not (false). A public project, such as the example-get-started demo project, can be accessed by anyone.
• team_name(string) - the name of the team associated with the project if we want to create the project in team space. This field is optional. If team name is not specified, the project will be in user namespace.
• subdir(string) - the subdirectory within the repository where the project will be located, in case of a monorepo. This field is optional.
• name (string) - the name of the project to be defined. If no name is specified, it will be extracted from the repository name. This field is optional.

Accepted formats for repo_url are:

• git@github.com:amritghimire/example-dvc-experiments.git
• https://github.com/nimdraugsael/example-dvc-experiments.git
• https://gitlab.example.org/amritghimire/example.git (For custom GitLab)
• ssh://git@gitlab.com/xxx/yyy/zzz/repo.git

Similar format is supported for GitHub, Bitbucket and GitLab.

Response

Here is an example JSON response from the API request:

{
  "projects": [
    {
      "id": "tsanhawn9y",
      "url": "https://studio.iterative.ai/user/amritghimire/projects/example-get-started-tsanhawn9y",
      "name": "example-get-started"
    }
  ]
}

This response confirms that the project was successfully created and provides the necessary information about the created project:

• id: The unique identifier of the project. In this case, the project ID is "tsanhawn9y".
• url: The URL of the created project. You can access the project using this URL. In this example, the project URL is https://studio.iterative.ai/user/amritghimire/projects/example-get-started-tsanhawn9y.
• name: The name of the created project. In this example, the project name is "example-get-started".

This information can be used to access and reference the project in the future.

Example

First, you need your DVC Studio client access token . For this example, we set it in the DVC_STUDIO_TOKEN environment variable:

$ export DVC_STUDIO_TOKEN=<TOKEN>

import json
import os
import requests


url = "https://studio.iterative.ai/api/create-project"
token = os.environ["DVC_STUDIO_TOKEN"]
payload = json.dumps({
  "repo_url": "https://github.com/amritghimire/monorepo-model-registry-fixture",
  "public": True,
  "team_name": "team_name",
  "subdir": "nested",
  "name": "test-nested"
})
headers = {
  'Authorization': f'token {token}',
  'Content-Type': 'application/json'
}

response = requests.request("POST", url, headers=headers, data=payload)
for project in response.json()["projects"]:
    print(project["url"])
    ...

$ curl --location 'https://studio.iterative.ai/api/create-project' \
--header "Authorization:token ${DVC_STUDIO_TOKEN}" \
--header 'Content-Type: application/json' \
--data '{
    "repo_url": "https://github.com/organization/repo",
    "public": true,
    "team_name": "team",
    "subdir": "nested",
    "name": "test-nested"
}'