House air waybill data extraction is the next document type coming to AWBGuru. Today AWBGuru reads master air waybills end to end. The new path will read the house air waybill (HAWB), the forwarder-issued document that describes one shipper's goods inside a consolidation, and turn it into a structured, validated house record that still points at its master.
If you are new to the idea, the guide to AWB data capture covers the basics. This post explains how HAWB data extraction is being built, and why it is not simply the master pipeline pointed at a different document. The short version: the design started from real house pages, and real house pages turned out to look very different from what most people picture when they hear "air waybill."
Most house bills don't look like an air waybill
The common assumption is that a house bill is a copy of the IATA air waybill form with a forwarder's own number in the corner. Some are. Most are not. In the set of 21 real house documents the classifier was tuned on, only 2 were printed on the IATA air waybill face. About 13 were forwarder layouts titled things like "Air Freight Manifest" or "Consolidation Manifest," each carrying a single house shipment. The remaining 6 or so were manifests carrying two or more houses on the same page.
That spread is the design problem. A tool that only understands the IATA form handles a small fraction of what cargo teams actually receive, and a tool that trusts the title on the page gets fooled, because a document called a "Consolidation Manifest" often describes exactly one house. We cover the different layouts in more detail in house air waybill formats, and the relationship between the two document levels in MAWB vs HAWB in a consolidation.
Classify first, then count the houses
Before any field is read, AWBGuru classifies each document into one of six types: master air waybill, house air waybill, consolidation manifest, commercial invoice, packing list, or other. Invoices and packing lists can arrive in the same email or PDF as the waybills, so knowing what each page is comes before deciding how to read it.
For house-level paperwork the deciding signal is not the title. It is the number of distinct house numbers on the page. The classifier notes the master AWB referenced in the header, lists every distinct house number, and then applies a simple rule:
- One house number: the page is a house air waybill, even when it is titled "Manifest" and shows a master number in its header.
- Two or more distinct house numbers: the page is a consolidation manifest, and it is classified as one and set aside rather than read as a single house bill.
- No house number: the page is not house-level paperwork at all and falls to the other types.
Counting instead of trusting the layout is what lets the same logic handle an IATA-form house bill, a one-house forwarder manifest and a multi-house consolidation manifest without a separate template for each.
With , master air waybill extraction runs on your own documents today, and house extraction will join the same pipeline.
Start hereA house schema built from real house pages
The master air waybill schema in AWBGuru has roughly 40 fields, because the IATA form has a box for almost everything, including the rate line, charges and declared values. The obvious shortcut for house bills would be to run them through that same schema. We looked at what real house pages actually carry before deciding against it.
A census of 75 real house pages recorded which fields were present on each one. The pattern is clear:
- On every page (100%): house AWB number, shipper, consignee, and nature of goods.
- On almost every page (99%): number of pieces and gross weight.
- 96%: the master AWB number, the link back to the consolidation.
- 85%: flight number. 81%: destination airport.
- Roughly three in four pages (75 to 77%): issuing forwarder, flight date, and origin airport.
Charges are close to absent, along with the fields around them. Freight charges and currency each appeared on 5% of pages, declared value on 7%, and chargeable weight on 9%. Those are exactly the fields that fill a large part of a master air waybill. Forcing the master schema onto a house bill would return a record that is mostly empty boxes, and it would ask reviewers to look at fields that were never on the page in the first place.
So the house schema will be leaner. It will be built around the fields that house pages reliably carry, with the master AWB number as a first-class field rather than an afterthought. A house record will look something like this:
The master link stays in the record
A house bill on its own answers "whose goods are these." It only becomes useful to a cargo team when it also answers "which master are they flying under." That is why the house record will keep the MAWB number alongside the house number. Every extracted house will trace back to its master, and houses that share a master can be grouped together again downstream, which is how consolidations are reconciled in the first place.
Because the master number is present on 96% of house pages, reading it is not a stretch goal. It will be part of the core record, and a house page that has no master reference is worth a second look by itself.
The same pipeline as master extraction
House air waybill data extraction is not a separate product bolted onto the side. It will run through the same intelligent document processing pipeline as master extraction, with the same three parts that make the output safe to act on:
- Per-field confidence: every field will come back with its own confidence, not one score for the whole page. See how confidence is calibrated.
- Validation: reads will be checked against air-freight logic before anything trusts them, because a confident read can still be wrong. More on that in validation vs confidence.
- Exceptions-only review: records that clear will go straight through, and a person will only see the fields that need one. That is the model described in exceptions-only review.
On the output side, house data will be produced as FHL, the IATA house message that sits alongside the FWB for the master; FHL vs FWB explains the difference.
Who this is for
House bills land on many desks, not just a forwarder's. Import and export cargo teams, handling agents, customs brokers preparing filings and operations staff reconciling a consolidation against its master all receive house-level paperwork in whatever layout the issuing forwarder happens to use. AWBGuru does not issue house bills. It will capture the ones you already receive and turn them into records your systems can use, so nobody has to key the same house number, shipper and weight from a PDF again.
For a product-level view of what the house path covers, see the HAWB data extraction page. For master documents, the full walkthrough is in how MAWB data extraction works.