Skip to content
NewTwenty now prepares Dubai Trade customs declarations →
Platform

Platform

OverviewHow it worksWorkbenchApprovalsMemoryMobileSecurity

Agents

SortQuoteBookReadClearTrack

Capabilities

InboxDocumentsSpot quotingRate requestsChasersCustoms prepHS codesPortalsReporting

Integrations

All integrationsEmailWhatsAppYour TMSCustoms portals
Customers

By business

ForwardersCustoms brokersNVOCCsBrokeragesCarriers

By team

PricingOperationsCustomsFinanceLeadership

By industry

Dangerous goodsCold chainProject cargoE commerceAutomotivePharmaAll industries

Use cases

QuotingBookingDeclarationsEmail triageTrackingInvoices

Proof

Case studiesMomentumForwarder inboxFreight networkDesign partners
Resources

Learn

BlogField notesGuidesUAE customsIncotermsGlossary

Watch

Twenty filmWalkthroughsFounder films

Free tools

ROI calculatorFreight rates

Product help

ChangelogHelp centerFAQFirst 30 daysNewsletter
Company

Company

AboutFoundersCareersNewsPartners

Talk to us

Book a sessionWhatsAppContactPress kit
Log inBook a working session
Field noteDocumentsOct 2026

Most shipping PDFs we receive have no text layer

In our test set of real shipping documents, nearly nine in ten PDFs were scans with no text to copy.

Twenty
Document bundleJ-22790 · 14 pages in one PDF
Commercial invoiceScan, no text layerRead as image
Packing list, 6 pagesScan, no text layerRead page by page
Bill of ladingScan, no text layerRead as image
Certificate of originDigital
Delivery orderScan, no text layerRead as image
Replay
5 documents found, 42 fields read

That rules out simple text extraction. Every page has to be read as an image.

Bundles often hold several documents in one file. Packing lists can run many pages. We split bundles first. Read long lists page by page and combine the results in code.

Document extraction

Bring one workflow.
Leave with a plan.

45 minutes on your real emails and documents.