Skip to main content
POST
/
extract
Create schema extraction job
curl --request POST \
  --url https://api.tile.run/v1/extract \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "file_id": "<string>",
  "document_schema": {}
}
'
{
  "extraction_id": "<string>",
  "status": "<string>",
  "created_at": "2023-11-07T05:31:56Z"
}

​
Document schema structure

The document_schema object should conform to the following structure. 
{
  "name": "Invoice",
  "description": "Schema for extracting invoice details",
  "fields": [
    {
      "name": "invoice_number",
      "description": "The unique identifier for the invoice",
      "type": "string"
    },
    {
      "name": "issue_date",
      "description": "The date when the invoice was issued",
      "type": "date"
    },
    {
      "name": "customer",
      "description": "Customer information",
      "type": "object",
      "fields": [
        {
          "name": "name",
          "description": "Customer's full name",
          "type": "string"
        },
        {
          "name": "email",
          "description": "Customer's email address",
          "type": "email"
        }
      ]
    },
    {
      "name": "line_items",
      "description": "List of items in the invoice",
      "type": "array",
      "fields": [
        {
          "name": "description",
          "description": "Item description",
          "type": "string"
        },
        {
          "name": "amount",
          "description": "Item cost",
          "type": "number"
        }
      ]
    }
  ]
}

​
Primitive fields

  • string: Generic text data
  • number: Numeric values
  • email: Email addresses
  • phone: Phone numbers
  • date: Date values

​
Objects and arrays

  • object: Nested object containing additional fields where each field is a primitive field.
  • array: List of items where each element is an object. As above, fields within each object can be any one of the primitive fields.

​
Best Practices

Field names

Use clear, descriptive names and avoid special characters. We recommend using snake case.

Descriptions

Descriptions are critical for accuracy. If you want to consistent formatting, include this in the description.

Nested structures

We support 3 levels of nesting for objects. Where possible, we recommend avoiding deeply nested objects as this reduces accuracy.

Nested objects within arrays

We do not currently support nested objects within arrays. This is on our roadmap.

Headers

​
Authorization
string<bearer>
required

Bearer token for authentication

Body

application/json
​
file_id
string
required

ID of the uploaded file to process

​
document_schema
object
required

Schema definition for extraction

Response

Schema extraction job created

​
extraction_id
string
required

ID of the created extraction job

​
status
string
required

Current status of the job

​
created_at
string<date-time>
required

When the job was created