This path shows how to capture invoice totals with the Extraction processor.

1. Upload and auto-detect

1. Go to https://app.algorythmos.fr and choose your workspace.
2. Click Upload and drag in an invoice PDF or image.
3. We auto-detect language, layout, and table presence in a few seconds.

2. Select the invoice template

• Choose Extraction → Invoices.
• Confirm key fields such as supplier_name, invoice_total, and due_date in the preview.
• Add custom fields on the right-hand panel; these become schema fields.

3. Review JSON output

{
  "supplier_name": "ACME Corp",
  "invoice_number": "INV-2048",
  "invoice_total": 582.33,
  "currency": "EUR",
  "due_date": "2025-10-15",
  "line_items": [
    { "description": "Support retainer", "quantity": 1, "unit_price": 582.33 }
  ]
}

4. Save as a processor

• Click Save Processor, name it invoice-extractor, and choose the Production workspace.
• Toggle Versioning → Auto publish for faster iteration.

5. Call the API

curl -X POST https://api.algorythmos.fr/processors/invoice-extractor/runs \
 -H "x-api-key: $ALG_KEY" \
 -H "Content-Type: application/json" \
 -d '{"input_path":"gs://algorythmos-datasets/invoices/acme.pdf"}'

The response includes a run_id. Poll GET /processors/invoice-extractor/runs/{run_id} to track status while you integrate the output into downstream systems.