Documents in. Clean data out.
Wove splits, classifies and extracts 35+ document types at 99.9% accuracy — PDFs, scans, spreadsheets and phone photos — cross-checks every field and sends clean data straight to CargoWise or your TMS.
Date 18 Sep 2026
Yangzi Home Textiles Co.
88 Jiangdong Rd, Ningbo
Pacific Home Goods Inc.
Los Angeles, CA
FOB Ningbo
T/T 30 days
8,240 KG
USD 184,206.31
35+ types, out of the box.
Any layout, any format — PDF, Excel, scans, phone photos and the email body itself.
From a messy inbox to clean data.
One PDF in. Every document identified.
Pre-alerts arrive as bundled PDFs, scans and email bodies. Wove splits them apart, recognises each document type and routes every page to the right place.
- Bundled packets split page by page
- Low-quality scans and phone photos handled
- The email body read as a document too
Every field and every line item.
Headers, parties, references and full line-item tables — even 2,000-line invoices — extracted with a confidence score on every value and matched to your CargoWise organisations.
- Rates, surcharges, HS codes and carrier-specific formats
- Parties matched to existing org codes
- Low-confidence fields flagged for review
Shanghai 200120, China
Mismatches caught before they reach your TMS.
Wove compares the same fields across every document in a shipment and flags anything that doesn't line up — with the values side by side and a plain-English reason.
- Weights, package counts, parties and containers
- Invoice and packing list totals reconciled
- Issues routed to the right person, not buried
Your documents, too.
Have a permit, a client-specific form or a local customs document? Define the fields you need with the Custom Document Types Builder — no engineering required.
- Point-and-click field definitions
- Validation rules on any field
- Synced like any built-in type
Clean data, wherever you need it.
Validated data syncs straight to CargoWise through the native eAdaptor integration, or to Magaya, Logiware or any system as structured JSON via REST API and webhooks.
Template-based OCR breaks every time a layout changes. See how LLM-native extraction compares.
Wove vs OCR →{
"document_type": "commercial_invoice",
"invoice_number": "INV-2026-04417",
"seller": { "name": "Yangzi Home Textiles Co.", "org_code": "YANHOMNGB" },
"buyer": { "name": "Pacific Home Goods Inc.", "org_code": "PACHOMLAX" },
"incoterms": "FOB",
"currency": "USD",
"total": 184206.31,
"gross_weight": { "value": 6180, "unit": "KG" },
"line_items": [
{
"description": "Cotton bed sheets, queen",
"hs_code": "6302.31",
"quantity": 2400,
"unit_price": 6.85,
"confidence": 0.998
}
],
"validation": { "status": "flagged", "issues": ["gross_weight_mismatch"] }
}
“We plugged Wove's document processing and HTS classification into our platform via API. Our customers get instant extraction and accurate classification on every shipment.”
ExFreightQuestions about document processing.
What is AI document processing for logistics?
It's software that reads shipping documents the way an experienced operator would — splitting bundled PDFs, recognising each document type, extracting every field and line item, and checking the data across documents — then sending structured data to your TMS instead of someone retyping it.
Which documents can Wove extract?
More than 35 logistics document types, including commercial invoices, packing lists, house and master bills of lading, air waybills, arrival notices, booking confirmations, certificates of origin, insurance certificates, fumigation certificates, ISF filings, CBP 7501 entry summaries, export declarations (AES/EEI), CMR road consignment notes, proofs of delivery, purchase orders, carrier invoices and rate sheets. You can also add your own types without engineering.
How accurate is Wove's extraction?
Wove's document extraction achieves 99.9% accuracy across 35+ logistics document types. Every field carries a confidence score, and low-confidence fields are flagged for a person to review instead of being passed through.
Does Wove need templates for each supplier or carrier?
No. Wove is LLM-native, so it reads new layouts, scans and phone photos without templates or training — unlike template-based OCR, which breaks whenever a document changes.
How does cross-document validation work?
Wove compares the same fields across every document in a shipment — consignee, package count, gross weight, container numbers, HS codes — and flags mismatches with the values side by side before anything reaches your TMS.
Where does the extracted data go?
Straight into CargoWise through the native eAdaptor integration, into Magaya or Logiware, or to any system as structured JSON via the REST API and webhooks.
Stop retyping documents.
Book a 20-minute demo and see Wove read, check and sync your shipping documents.