Scan a receipt or invoice into a proposed voucher

Takes a receipt or invoice as RAW BYTES — a PDF, an image or plain text, uploaded under its own content type, not wrapped in JSON — extracts what the…

POST /v1/books/scan

Books → · All API references → · Six flows, four surfaces →


Address	`https://api.hanzo.ai/v1/books/scan`
Method	`POST`
Operation	`post_books_scan`
Auth	`Authorization: Bearer $HANZO_API_KEY`

Takes a receipt or invoice as RAW BYTES — a PDF, an image or plain text, uploaded under its own content type, not wrapped in JSON — extracts what the document says, resolves the vendor's expense category, and answers a DRAFT carrying a balanced voucher proposed for it.

NOTHING IS POSTED. That split is the whole design: the model only ever produces a structured reading of the document, the voucher is assembled deterministically in Go from that reading, and the ledger is written only by the separate book call a human confirms. So a misread scan can propose a wrong draft; it cannot move money. Amounts are exact integer cents end to end — the extraction returns cents, never a decimal — so no rounding enters the ledger.

The draft's id is the FILE HASH, and that is what makes booking idempotent: re-scanning the same bytes addresses the same draft rather than queuing a second one. A row is written to the org's document inbox as a side effect, moving it from unsorted to draft. Scoped to the caller's own org from the validated principal and refused without one; sandbox=true targets the sandbox ledger, and filename is recorded for the inbox. An empty or oversized upload is a 400, and a deployment with no scanner model answers 501.

Request

1 field, body application/octet-stream.

Field	In	Type	Required	Description
`(body)`	body	string (binary)	yes

Response

Status	Body	Meaning
`2XX`	ScanDraft	Success

2XX body — 32 fields.

Field	In	Type	Always	Description
`balanced`	body	boolean	—
`category`	body	string	—
`confidence`	body	string	—
`extracted`	body	Extracted	—
`extracted.category`	body	string	—	Category is the expense bucket the SCANNER guessed, as a slug — a hint only.
`extracted.currency`	body	string	—	Currency is the ISO code the document is denominated in.
`extracted.issuedAt`	body	string	—	IssuedAt is the document's OWN date as YYYY-MM-DD — when the bill was issued, which is not when it was uploaded or when it will post.
`extracted.lineItems`	body	LineItem[]	—	LineItems are the individual lines read off the document, where it had any.
`extracted.lineItems[].amountCents`	body	integer	—	AmountCents is that line's amount in whole cents.
`extracted.lineItems[].description`	body	string	—	Description is the line as it appears on the document.
`extracted.merchant`	body	string	—	Merchant is the supplier as printed on the document.
`extracted.note`	body	string	—	Note is anything else worth carrying from the document that has no field of its own.
`extracted.taxCents`	body	integer	—	TaxCents is how much of that total is tax, in cents.
`extracted.totalCents`	body	integer	—	TotalCents is the document total in whole cents, tax INCLUDED.
`questions`	body	Question[]	—
`questions[].account`	body	string	—	Account is the chart number the questioned entry posted to, where one applies.
`questions[].amount`	body	string	—	Amount is the figure that makes the question concrete, already FORMATTED for display with its currency symbol — a string, not cents, and not for arithmetic.
`questions[].id`	body	string	—	ID is the source transaction the question is about, so answering it leads straight back to the entry that raised it.
`questions[].kind`	body	string	—	Kind is what looked wrong: outlier (a charge far above the usual), reversal (a posting undone), roundoff (a balancing plug big enough to be worth explaining),…
`questions[].postedAt`	body	string	—	PostedAt anchors the question in time — when the entry it concerns posted.
`questions[].text`	body	string	—	Text is the question itself, written for a founder to answer directly.
`scanId`	body	string	—
`vendor`	body	string	—
`voucher`	body	Voucher	—
`voucher.description`	body	string	—	Description is the human line for the event, e.g.
`voucher.legs`	body	Leg[]	—	Legs are the sides of the posting.
`voucher.legs[].account`	body	string	—	Account is the chart-of-accounts number this side posts to, e.g.
`voucher.legs[].credit`	body	integer	—	Credit is the leg's credit in exact cents.
`voucher.legs[].debit`	body	integer	—	Debit is the leg's debit in exact cents.
`voucher.postingAt`	body	string	—	PostingAt is the RFC3339 instant the event posts at — the time every statement window filters on.
`voucher.sourceId`	body	string	—	SourceID is the source event's own id within that namespace.
`voucher.sourceKind`	body	string	—	SourceKind is the idempotency namespace naming what booked this, e.g.

Failure carries the platform error shape — see Errors.

Examples

hanzo books scan create

import { Configuration, BooksApi } from 'hanzoai';

const api = new BooksApi(new Configuration({ accessToken: process.env.HANZO_API_KEY }));
const { data } = await api.postBooksScan();

from hanzoai.cloud import ApiClient, Configuration
from hanzoai.cloud.api import BooksApi

client = ApiClient(Configuration(access_token=os.environ["HANZO_API_KEY"]))
result = BooksApi(client).post_books_scan()

cfg := cloud.NewConfiguration()
cfg.AddDefaultHeader("Authorization", "Bearer "+os.Getenv("HANZO_API_KEY"))
client := cloud.NewAPIClient(cfg)

resp, _, err := client.BooksAPI.PostBooksScan(context.Background()).Execute()
if err != nil {
	return err
}

use hanzo_cloud::apis::{configuration::Configuration, books_api};

let mut cfg = Configuration::new();
cfg.bearer_access_token = std::env::var("HANZO_API_KEY").ok();

let result = books_api::post_books_scan(&cfg, Default::default()).await?;

import ai.hanzo.cloud.ApiClient;
import ai.hanzo.cloud.api.BooksApi;

ApiClient client = new ApiClient();
client.setRequestInterceptor(b -> b.header("Authorization", "Bearer " + System.getenv("HANZO_API_KEY")));

var result = new BooksApi(client).postBooksScan();

curl -X POST https://api.hanzo.ai/v1/books/scan \
  -H "Authorization: Bearer $HANZO_API_KEY"

The door reaches books through the books tool, which names its 24 operations with its own verbs — this one among them, under a name only the door declares. describe explains any of them:

curl -X POST https://api.hanzo.ai/v1/mcp \
  -H "Content-Type: application/json" \
  -d '{
       "jsonrpc": "2.0",
       "id": 1,
       "method": "tools/call",
       "params": {
         "name": "describe",
         "arguments": {
           "op": "list_book_accounts"
         }
       }
     }'

Books API · All Hanzo APIs · Interactive reference

Scan a receipt or invoice into a proposed voucher

Request

Response

Examples

On this page