Your BOLs, rate confirmations and freight invoices, as a spreadsheet.
Upload up to 30 documents. Get back one CSV, one row per document, matched to your file names, with low-confidence fields flagged for your review. No call, no demo, no contract. The first 30 are free.
What you get
A CSV you can open in Excel or Google Sheets, or import into most TMS platforms. One row per document. The column set is fixed per document type, so a batch from next month lines up with a batch from this month.
Bill of Lading
- bol_number
- bol_date
- shipper_name
- shipper_address
- consignee_name
- consignee_address
- carrier_name
- po_number
- pro_number
- piece_count
- weight_lbs
- commodity_description
Rate Confirmation
- rc_number
- load_number
- broker_name
- carrier_name
- carrier_mc_number
- rate_total_usd
- origin_city_state
- destination_city_state
- pickup_date
- delivery_date
- equipment_type
- weight_lbs
- commodity_description
Freight Invoice
- invoice_number
- invoice_date
- due_date
- carrier_name
- bill_to_name
- load_number
- pro_number
- amount_total_usd
- line_haul_usd
- accessorials_usd
- payment_terms
- remit_to
Every row also carries file_name (your original file name, for audit trail), doc_type, and a low_confidence_fields column listing the fields we were not sure about. We never auto-correct a value. A field that reads ambiguously stays as read and gets flagged, so you check it against the original instead of trusting a number you cannot verify.
How it works
Up to 30 PDFs or images (JPG, PNG), 8 MB each. Pick the document type, or "mixed" and we sort them.
Each document is read by the same AI vision pipeline that ships in Vantage YMS, with a fixed field set and a confidence score per field.
A download link lands in your inbox, typically within minutes and within 48 hours at most. Files are deleted after 30 days.
Send 30 documents. No account.
Free, one time per company. Results go to the email you enter. If you would rather email the files, send them to hello@vantageyms.com with "FreightSheet pilot" in the subject.
Pricing, in the open
Most document-extraction vendors make you book a call to find out what this costs. We do not think that is necessary.
- Spreadsheet returned within 48 hours
- Low-confidence flags included
- No card, no account, no call
- Up to 500 documents per month
- 24-hour turnaround on business days
- Email support, answered in writing
- Month to month, cancel with one month's notice
- Applies above 500 documents in a month
- Or a custom monthly rate at 1,000+
- Priority processing
| Rossum | Veryfi | Tai TMS (broker back office) | FreightSheet | |
|---|---|---|---|---|
| Starting price | $18,000/year | $500/month minimum | $995 to $7,925/month | $1,500/month |
| Contract | Annual | Monthly | Annual (typical) | Month to month |
| Setup fee | Not published (sales call required) | Not published | Not published | $0 |
| Free trial | Demo call | Demo call | Demo call | 30 documents, free, no call |
Sources: Rossum, Veryfi and Tai TMS published pricing pages as of this writing. Payment for the retainer is monthly in advance by card or bank transfer.
How we measure accuracy (a method, not a claim)
We are not going to tell you "99% accurate" without showing our work. Most vendors in this space do, and it is usually not verifiable. Here is how we measure it, and what we will show once there is real production volume behind it:
- Field-level comparison. Every extracted field is compared against a human-verified ground truth for a held-out sample of documents.
- Two numbers, not one. Exact match rate on structured fields (dates, dollar amounts, reference numbers), and match rate after normalizing formatting differences ("01/15/26" and "Jan 15 2026" count as a match).
- Confidence flagging is separate from accuracy. Fields we flag as low-confidence are excluded from the headline number and reported separately, so the number is not inflated by hiding uncertain fields.
- Per-document-type breakdown. BOL, rate confirmation and invoice results are reported separately. They are different documents with different failure modes.
We would rather publish a real number in a few weeks than a made-up one today. If you send your 30 documents, your batch becomes part of that first real number.
What this is not
Fixed price only works with a fixed scope. This is the whole list, and it is the same list in the email and the agreement.
- No TMS or ERP integration. No API connection. File upload or email attachment in, spreadsheet out.
- Printed or scanned English documents first. Handwriting is best effort and will carry low-confidence flags.
- English only for now.
- No data correction. We do not decide a value is wrong and fix it. We flag it.
- Not legal or accounting advice. Decisions made on the extracted numbers are yours.
- Batch, not real time. 24 to 48 hours, not a live API.
- Not unlimited. 500 documents per month on the retainer; above that is billed per document.
Questions, answered in writing
Is our data secure?
Documents go into a private storage bucket with no public access, are processed, and are deleted on a 30-day schedule. We do not share your documents or extracted data with any third party, and we do not use your documents to train models for other customers. If you need a signed data processing agreement before sending documents, ask and we will send one. No call required.
How accurate is it, really?
We publish our accuracy methodology openly (above) rather than a single marketing number. Every batch comes back with low-confidence fields flagged separately, so you are never trusting a number you cannot verify. You can spot-check any flagged field against the original document yourself.
Does this connect to our TMS?
Not directly. This is a file-in, spreadsheet-out service, not a software integration. You upload documents, we return a spreadsheet you can import into whatever system you use (most TMS platforms accept CSV import). If direct integration becomes something several customers need, we will build it, but the core service works without any technical setup on your end.
Why is this so much cheaper than Rossum, Veryfi or Tai?
Those platforms are built for enterprise-scale automated pipelines with dedicated account teams and long-term contracts, and that overhead is priced in. We are a smaller, focused operation serving brokers who need accurate extraction without an enterprise contract. You get the extraction, not the sales team or the annual commitment. If your volume grows past what a lean setup like ours can handle, we will tell you.
You are a small operation. Can I trust this?
Fair question. Here is what is verifiable: the extraction runs on the same OCR and AI vision pipeline that ships in Vantage YMS on the App Store, where it reads trailer and container numbers at warehouse gates and validates them against ISO 6346 checksums in production. Beyond that, the fastest way to build trust is to test it: send 30 documents and judge the results directly rather than taking anyone's word for it.
Who owns the data, ours or yours?
Your documents and the extracted data are yours. We do not claim ownership, we do not resell it, and we do not use it to train shared models across customers. Once we deliver your spreadsheet, source documents are deleted from our processing storage on a routine 30-day schedule.
What are the contract terms?
Month to month. No annual commitment, no cancellation fee. If the pilot works for you, the retainer starts at $1,500 per month for up to 500 documents; if you need to stop, one month's notice by email is all it takes.
Can we really do this without a call?
Yes. That is the point. Everything you need to decide is on this page: what is included, what it costs, what happens to your data, and how to test it for free. If a question comes up that is not answered here, email us and we will answer it in writing. A phone call is not a requirement to start or to buy.