JSON schema: extract a U.S. invoice

JSON Schema to extract a U.S. invoice with an LLM: vendor, invoice number, dates, line items, subtotal, sales tax and total in USD.

Extract the invoice data. Amounts are plain numbers in U.S. dollars, without the $ sign. If a field is missing, leave it out: never invent it.

JSON Schema

{
  "type": "object",
  "required": [
    "vendor",
    "invoice_date",
    "total"
  ],
  "properties": {
    "total": {
      "type": "number"
    },
    "vendor": {
      "type": "string",
      "description": "Legal name of the vendor"
    },
    "currency": {
      "enum": [
        "USD"
      ],
      "type": "string"
    },
    "due_date": {
      "type": "string",
      "format": "date"
    },
    "subtotal": {
      "type": "number"
    },
    "sales_tax": {
      "type": "number",
      "description": "Sales tax amount, 0 if none"
    },
    "line_items": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "description",
          "amount"
        ],
        "properties": {
          "amount": {
            "type": "number"
          },
          "quantity": {
            "type": "number"
          },
          "unit_price": {
            "type": "number"
          },
          "description": {
            "type": "string"
          }
        }
      }
    },
    "invoice_date": {
      "type": "string",
      "format": "date",
      "description": "Issue date (YYYY-MM-DD)"
    },
    "invoice_number": {
      "type": "string"
    }
  }
}

Ejemplo de salida

{
  "vendor": "Bright Print Co.",
  "invoice_number": "INV-2026-0412",
  "invoice_date": "2026-09-15",
  "due_date": "2026-10-15",
  "line_items": [
    {
      "description": "Business cards x500",
      "quantity": 1,
      "unit_price": 80.0,
      "amount": 80.0
    },
    {
      "description": "Flyers 5.5x8.5 x1000",
      "quantity": 1,
      "unit_price": 120.0,
      "amount": 120.0
    }
  ],
  "subtotal": 200.0,
  "sales_tax": 17.25,
  "total": 217.25,
  "currency": "USD"
}

Coste estimado

Tokens de entrada por llamada : ≈ 656

Solo coste de entrada, sin la respuesta del modelo. Precios públicos estándar de los proveedores (Mistral AI, DeepSeek, OpenAI, Anthropic) convertidos de USD a euros al tipo del BCE del 29 de septiembre de 2026. Estimación: 1 token ≈ 3,6 caracteres.

Identificador

contextetech--extract-us-invoice

Tipo

Esquemas de salida

Archivo

extract-us-invoice.schema.json

Formato

JSON (application/json)

Tamaño

1215 bytes (1,2 KB)

Codificación

UTF-8

Tokens estimados

≈ 338

Idioma

Inglés

Licencia

MIT

Uso comercial

Permitido

Datos personales

Ninguno

Versión

1.0

Publicado el

2 de octubre de 2026 a las 21:54

Actualizado el

2 de octubre de 2026 a las 21:54

Almacenamiento

en base de datos

Usos

0

Me gusta

0

extract-us-invoice.schema.json

{
  "type": "object",
  "required": [
    "vendor",
    "invoice_date",
    "total"
  ],
  "properties": {
    "total": {
      "type": "number"
    },
    "vendor": {
      "type": "string",
      "description": "Legal name of the vendor"
    },
    "currency": {
      "enum": [
        "USD"
      ],
      "type": "string"
    },
    "due_date": {
      "type": "string",
      "format": "date"
    },
    "subtotal": {
      "type": "number"
    },
    "sales_tax": {
      "type": "number",
      "description": "Sales tax amount, 0 if none"
    },
    "line_items": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "description",
          "amount"
        ],
        "properties": {
          "amount": {
            "type": "number"
          },
          "quantity": {
            "type": "number"
          },
          "unit_price": {
            "type": "number"
          },
          "description": {
            "type": "string"
          }
        }
      }
    },
    "invoice_date": {
      "type": "string",
      "format": "date",
      "description": "Issue date (YYYY-MM-DD)"
    },
    "invoice_number": {
      "type": "string"
    }
  }
}

API de Anthropic (Python)

import json, anthropic

schema = json.load(open("extract-us-invoice.schema.json"))
client = anthropic.Anthropic()
resp = client.messages.create(model="claude-sonnet-5", max_tokens=1024,
    tools=[{"name": "extraire", "description": "Renvoie les données extraites", "input_schema": schema}],
    tool_choice={"type": "tool", "name": "extraire"},
    messages=[{"role": "user", "content": "tu entrada aquí"}])
print(resp.content[0].input)

Formato OpenAI (response_format)

{
  "type": "json_schema",
  "json_schema": {
    "name": "extract-us-invoice",
    "schema": {
      "type": "object",
      "required": [
        "vendor",
        "invoice_date",
        "total"
      ],
      "properties": {
        "total": {
          "type": "number"
        },
        "vendor": {
          "type": "string",
          "description": "Legal name of the vendor"
        },
        "currency": {
          "enum": [
            "USD"
          ],
          "type": "string"
        },
        "due_date": {
          "type": "string",
          "format": "date"
        },
        "subtotal": {
          "type": "number"
        },
        "sales_tax": {
          "type": "number",
          "description": "Sales tax amount, 0 if none"
        },
        "line_items": {
          "type": "array",
          "items": {
            "type": "object",
            "required": [
              "description",
              "amount"
            ],
            "properties": {
              "amount": {
                "type": "number"
              },
              "quantity": {
                "type": "number"
              },
              "unit_price": {
                "type": "number"
              },
              "description": {
                "type": "string"
              }
            }
          }
        },
        "invoice_date": {
          "type": "string",
          "format": "date",
          "description": "Issue date (YYYY-MM-DD)"
        },
        "invoice_number": {
          "type": "string"
        }
      }
    },
    "strict": true
  }
}