Built by an ex-WiseTech Global engineer
From messy email packets to filing-ready customs entries in minutes
Upload the invoice, packing list, B/L or AWB and certificate of origin. EntryWrite extracts, validates and maps 50+ fields to your entry system, ready for ACE/ABI filing with an audit trail.
No card, no integration work to start. Your first 20 shipments are on us.
Mapped to the entry systems small brokers actually file in
- CargoWise
- Descartes
- NetCHB
- Magaya
- QP/WP
The cost of the status quo
The re-keying already has a price. It is just buried in a salary.
EntryWrite has not run its first pilot yet, so there are no customer results on this page. These are the numbers that describe the problem today, and what the pilot is aiming at.
- Industry benchmark
45-60min
Manual keying per US import entry
The typical hands-on-keyboard time to key one entry from a document packet today, before anyone checks it.
Industry benchmark - Industry benchmark
$24-36/hr
Advertised rate for a customs entry writer
Taken from live US job postings for "Customs Entry Writer" and "Import/Export Clerk" roles. This is the cost the re-keying currently sits on.
- Pilot target
60-80%
Straight-through processing
What the pilot aims for after two to three weeks of rule tuning on your lanes. Stated as a target, not a measured customer result.
How it works
Four steps, and a reason each one matters to you
Extraction on its own is commoditised. What removes hours from someone's week is what happens on either side of it.
Ingest the whole packet
Forward the email, drop the PDFs, or connect an inbox or shared folder. One upload per shipment - invoice, packing list, B/L or AWB, and certificate of origin together.
No new system to learn
Your team keeps working the way it already does. Packets arrive by email today, so EntryWrite reads them where they already land instead of asking anyone to change their process.
Extract and reconcile 50+ fields
Importer and consignee, MID, seller, HTS, value, freight, insurance, origin, container references and line items - pulled from every document in the packet and reconciled against each other.
One shipment, not four documents
The hard part was never reading a single invoice. It is deciding which of four documents is right when they disagree. EntryWrite reconciles the packet and shows you where the conflict is.
Validate, then review only exceptions
Rules check invoice totals against line totals, weights and package counts across PL and B/L, HTS against description, and PGA data completeness. Low-confidence fields are flagged.
Your team touches the 20% that needs a human
Clean shipments pass straight through. People spend their time on genuine mismatches instead of re-typing fields that were already correct.
Export filing-ready data
Push mapped entries into your brokerage system via API, SFTP or CSV, or download ACE-ready files.
Mapped to your schema, not a generic spreadsheet
Output lands in the field names your entry system expects, with field-level confidence scores and every value traced back to its source page for CF-28 and post-entry audits.
Why this is not just OCR
The defensible part is everything around the extraction
Any model can read an invoice. Very few things can tell you that the invoice and the packing list disagree, which of the two your entry system should believe, and prove where the number came from eighteen months later.
Mapped to your entry schema
Pre-built field mappings for the systems small brokers actually file in - CargoWise, Descartes, NetCHB, Magaya, QP/WP - so output arrives in the field names your system expects.
Validation, not just extraction
Cross-document checks between invoice, packing list and B/L, HTS-against-description sanity checks, PGA data completeness, and low-confidence flags on every uncertain field.
Exception-first workflow
Humans review only mismatches and missing data. Clean shipments process straight through, so the queue in front of your team is the short one.
Audit-ready by design
Every field traces back to its exact source page, with immutable logs built for CF-28 responses and post-entry audits. Defensibility is a feature, not an afterthought.
Capabilities
What you get on day one
Everything below is aimed at one number: how many of your shipments get through without a person touching them.
Packet-level ingestion
One upload per shipment: commercial invoice, packing list, bill of lading or air waybill, and certificate of origin. Combined PDFs are split automatically.
50+ fields per entry
Importer and consignee, MID, seller, HTS, value, freight, insurance, origin, container references, line items and more - extracted per entry, not per page.
Cross-document validation
Totals match line totals. Weights and package counts align across packing list and B/L. HTS codes are checked against descriptions. Missing PGA data raises an alert.
Schema mapping to your system
Pre-built mappings for CargoWise, Descartes, NetCHB, Magaya and QP/WP, with a field-level confidence score on every mapped value.
Exception queue
Review only low-confidence fields and cross-document mismatches. Approve or correct in a few clicks, with the source document open beside the field.
Batch processing
Upload 50 to 200 shipments at once and get reconciled, validated entry drafts back in minutes rather than over an afternoon.
Audit trail and provenance
Click any field to see the exact source page it came from. Export complete entry packets when CBP asks.
Integrations and exports
Push to your brokerage software or TMS by API, SFTP or CSV. ACE/ABI-ready output formats.
Security and compliance
Role-based access, SSO, encryption in transit and at rest, and immutable logs for compliance reviews.
Free pilot
Your first 20 shipments, processed end to end
Send real import packets from your own lanes. We map them to your entry schema, run the validation rules, and show you how many went straight through. No obligation, no card, no integration work required to start.
Free pilot
20 shipments, free
Run on your own packets. Nothing to install, nothing to sign beyond a DPA.
- 20 shipments processed end to end
- Field mapping configured for your entry system
- Validation rules tuned to your lanes
- A straight-through-processing report on your real packets
- A data processing agreement before you send anything
After the pilot
Transparent per-shipment or tiered pricing, sized to your volume. No quote-only pricing games, and no enterprise floor that a five-person brokerage cannot justify.
- Priced for small brokerages
- Built for 3 to 25 person independent brokers and forwarders, not for an enterprise procurement cycle.
- No per-seat tax on reviewers
- Exception review is the part that keeps you compliant. Charging per reviewer would penalise the workflow that makes the product safe.
- You keep your data
- Export everything at any time. No lock-in through your own entry history.
We will publish per-shipment pricing once the first pilots tell us the right shape for it.
FAQ
The questions brokers actually ask
Is this just OCR?
No. Raw extraction is commoditised - a general model can read an invoice. The work that actually costs your team time is mapping fields to your entry schema, validating them across four documents that disagree, queueing the exceptions for a person, and keeping an audit trail. That is what EntryWrite does.
Do you file directly to ACE/ABI?
We produce filing-ready, mapped entry data and can integrate with ABI-connected systems. Some partners file directly through our API. EntryWrite is not itself the filing system - it removes the keying that happens before you file.
What document types do you support?
Commercial invoices, packing lists, bills of lading and air waybills, certificates of origin, arrival notices, and common PGA forms. Combined PDFs containing several of these are split automatically.
How long does setup take?
Most brokers are processing live shipments within days using email or folder intake. Deeper system integrations follow once the extraction and validation rules are tuned to your lanes.
Who built this?
EntryWrite is built by Ariel Serravalle, an engineer who worked at WiseTech Global - the company behind CargoWise. That means the field mappings are designed by someone who has worked inside the software you file in, rather than inferred from documentation.
What happens to our documents?
Documents are encrypted in transit and at rest, processed only to produce your entry data, and never used to train models. See the security page for the full detail, and ask for a data processing agreement before you send anything real.
Stop re-keying 50+ fields per shipment
Run a no-obligation pilot on your real import packets and see how many entries go straight through.
- 01You forward one shipment packet from a real lane.
- 02We map it to your entry schema and run the validation rules.
- 03You see exactly which fields came through clean, and which needed a human.