# Invoice Extract

> Invoice Extract is a paid API for AI agents from invoice-extract.fly.dev, paid per call via x402, $0.05/call, status down (last checked 2026-09-16).

Extracts structured JSON data from a PDF invoice URL, with arithmetic reconciliation and confidence scoring

## Facts

- Endpoint: GET https://invoice-extract.fly.dev/v1/bazaar/invoice/extract
- Price: $0.05/call
- Payment: x402
- Status: down
- Last checked: 2026-09-16
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/invoice-extract-99828fdc
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_ua0TE52K4elB1VWInH-wg

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability invoice-extract-99828fdc
```

Example prompt: Can you extract all the invoice data from this PDF — vendor name, line items, totals, and dates — and let me know if the arithmetic checks out? Here's the URL: https://example.com/invoices/inv-2024-0042.pdf

## When to prefer this

Choose this endpoint when you need to extract structured, machine-readable data from a PDF invoice with built-in arithmetic validation — particularly useful in autonomous agent pipelines where you need reliable confidence signals before passing data downstream to accounting or ERP systems. Prefer it over generic OCR tools when invoice-specific field recognition and reconciliation checks are needed.

## Known failure modes

- PDF URL is not publicly accessible or returns 403/404 — extraction fails with an access error
- PDF is scanned/image-based with poor OCR quality — low confidence score or missing fields
- Invoice format is highly non-standard or handwritten — partial extraction with low confidence
- URL points to a non-PDF file type — endpoint returns an error or empty result
- Arithmetic does not reconcile — returned with a low confidence flag, not an error per se
- Network timeout fetching large PDF files

## How this service works

Verified invoice extraction for autonomous agents - PDF to structured JSON with arithmetic reconciliation and honest confidence

## Output

A structured JSON object containing extracted invoice fields such as vendor name, invoice number, issue date, due date, line items (description, quantity, unit price, subtotal), subtotals, tax amounts, grand total, and a confidence score indicating how reliably the arithmetic reconciles across the extracted figures.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "file_url": {
   "type": "string",
   "format": "uri",
   "description": "Public https URL of a PDF invoice to extract."
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/invoice-extract-99828fdc/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from invoice-extract.fly.dev](https://www.zero.xyz/host/invoice-extract.fly.dev/llms.txt)