Documentation

File storage and loading data

4. File storage

Every project has file storage, used as a source for data loads and for dashboard images. Requires MD Admin.

REST

Files are addressed by fileId, which is the URL-safe base64 of "<name>|<parentPath>".

Method	Path	Description
GET	`/projects/{project}/storage/list?withStats=1`	List the root folder
POST	`/projects/{project}/storage/list`	Create a folder: `{ "folder": "/imports/2026" }`
GET	`/projects/{project}/storage/list/{fileId}`	File or folder details and contents
DELETE	`/projects/{project}/storage/list/{fileId}`	Delete a file, or a folder recursively
GET	`/projects/{project}/storage/list/{fileId}/download`	Download a file
POST	`/projects/{project}/storage/move`	Move or rename: `{ "source": "/a.csv", "target": "/archive/a.csv" }`
POST	`/projects/{project}/storage/upload`	Start a chunked upload (resumable.js protocol)

WebDAV

For simple file transfer, the storage is also available over WebDAV, so any WebDAV client or a mounted drive works:

https://app.inzata.com/api/v1/projects/{project}/webdav/

Authentication is HTTP Basic: username = your user id, password = your project WebDAV token. A project Admin sets the token.

BASH

curl -u "[email protected]:<webdav-token>" \
     -T sales_2026.csv \
     https://app.inzata.com/api/v1/projects/{project}/webdav/imports/sales_2026.csv

Uploading a file can also start a data load automatically (see the WebDAV file-change trigger in section 6).

5. Loading data

From a file (upload → guess → load)

1. Upload the file to storage (REST upload or WebDAV).

2. Detect the structure. Inzata guesses columns, data types and keys:

HTTP

GET /api/v1/projects/{project}/data/guessupload/{uploadId}

JSON

{
  "ident": "sales_2026",
  "prescription": {
    "clustername": "sales",
    "primarykey": "order_id",
    "columns": [
      { "id": "order_id", "originalheader": "Order ID", "objectType": "A", "dataType": "string", "label": "Order ID" },
      { "id": "amount",   "originalheader": "Amount",   "objectType": "F", "dataType": "number", "label": "Amount" }
    ],
    "data": [["1001", "250.00"], ["1002", "99.90"]]
  }
}

objectType is A (attribute), L (label) or F (fact).

3. Adjust the structure if needed: PUT /data/guessupload/{uploadId} with the edited prescription.

4. Load the data:

HTTP

POST /api/v1/projects/{project}/kodiak/transform/execute
Content-Type: application/json

{ "type": "cluster", "clusters": { "sales": { "source_type": "webdav", "source": "/imports/sales_2026.csv" } } }

The response contains an scache job link. Poll it until it finishes (see Asynchronous jobs).

Other resources

Method	Path	Description
GET	`/projects/{project}/kodiak/dataload`	Current data and model version
GET	`/projects/{project}/kodiak/scache/obj?type=transform&status=ERROR`	List jobs, filtered by type (`transform`, `flow`, `report`, `ai`) and status
GET	`/projects/{project}/store/package`	Prebuilt data packages available for the project
POST	`/projects/{project}/package/obj`	Load a package and join it to your data