Wove vs Ripple for Forwarder Customs Workflow 2026

The filing decision, not a logo contest

Customs workflow automation for a freight forwarder is the path that turns a commercial invoice, a packing list, and a bill of lading into a checked entry draft inside the system the broker already files from. The person who typed this search is close to a choice. The choice is which product prepares that draft, which system of record receives it, and who still presses the filing click. It is not a beauty contest between two home pages.

Wove, on the pages opened for this guide, sells an execution layer for global trade. It reads shared-inbox mail and shipping documents, creates shipments and one-off quotes, builds customs entry drafts, matches carrier invoices, and publishes a tariff calculator plus a tariff API. The forwarder and broker pages center on CargoWise, with Magaya, Logiware, and Oracle NetSuite also named. Ripple, the London logistics company at useripple.io, sells agents that run operational work across email, documents, and the systems a logistics team already has. Its customs page is built around UK clearance, and its wider site names classification, quote-to-cash, vendor control, audits, incidents, and ESG reporting. A request to useripple.com did not return a page in this research session. The product that loaded is the logistics company at useripple.io. Nothing on the pages opened shows a rename, a merger, or a shutdown of either product.

A company directory lists this Ripple as founded in 2024 in London, according to CB Insights (2026). Treat that as a directory listing, not a regulatory filing. Wove's chief executive told trade press the company is a technology vendor and is not itself a forwarder or a customs broker. That distinction matters, because the license to file stays with a person either way.

The short version is this. Pick Wove when the desk already lives in CargoWise and tariff history is part of the buy. Pick Ripple when the pain is messy customs prep, especially into a UK declaration system, or when the same vendor is also meant to take on quote-to-cash and compliance work beside the entry. Then confirm, in writing, who still files.

2,000-line invoice: about 30 seconds according to FreightWaves (2026), which also reports about 3.2 million HTS rules across more than 45 revisions, a 431-page notice ingested in about one hour, fundraising under $500,000, and sevenfold annual recurring revenue growth over the past year. Those figures are the vendor's account to a reporter, not a lab result from this guide.

Key Takeaways

• The buy is about where the draft is written and who still files, not about which logo sounds more automatic.

• Wove's public product is inbox-to-shipment, entry drafts, quotes, payables match, and a tariff data layer, with CargoWise as the native write-back.

• Ripple's public product is customized operational execution, with a detailed UK customs-prep story and a wider compliance brief.

• Neither vendor published a dollar price on the pages opened. Both are Quote-based until a buyer gets a written offer.

• A weight break, a missing origin, and the filing click still need a named person.

• A mailbox-to-draft layer can sit above either tool only when the API or the export exists and a reviewer is assigned before anyone talks about going live.

Who this is for

This guide is for an operations lead at a freight forwarder or a customs broker who already has a system of record and is deciding whether Wove or Ripple should prepare the customs paperwork around it. It also fits a branch manager who is tired of re-keying commercial invoices and wants to know what remains a human job after the software is in. It is written for someone who will sit in a scoping call this month, not for a student surveying the whole software market.

You are in the right place if your team receives pre-alerts by email, classifies lines, and then rebuilds the same facts inside CargoWise, Sequoia, or another customs system. You are also in the right place if tariff changes keep landing after the quote has gone out and you need a product that can show its work. A desk that only wants a generic chatbot over a shared drive will not get a decision from this comparison.

Red flags: you want software to replace the licensed filer; you require a public price list before any conversation; your system of record is one neither vendor has shown a write-back for, and you will not fund an export or an API step.

Use this checklist before you book time with either vendor. Name the system that must receive the draft. Name the country or countries you file in this quarter. Name the document types that arrive broken, scanned, or merged. Name the person who may reject a line. If you cannot fill those four blanks, you are not ready to compare products, and a pilot will turn into a second inbox.

How we evaluated these tools

The weighting below is a buyer frame for this reader. It is not a score we awarded to either vendor. Several facts are simply not on the public pages, and inventing a point total would hide that. Document intake from email packs carries 20 percent because that is where the forwarder's day actually starts. Classification and duty logic carry 20 percent because a wrong code becomes a delay, a bill, or a penalty. Write-back to the system of record carries 20 percent because the labor the team wants to stop is the re-key into CargoWise or the declaration system. Exception routing and the audit trail carry 15 percent because the broker's judgment sits on the lines the software should refuse to guess. Filing geography carries 15 percent because a UK Customs Declaration Service desk and a US consumption-entry desk do not have the same Tuesday. Price clarity and time to a first live file carry 10 percent because both products are Quote-based, so a public sticker cannot dominate the decision, while the shape of the rollout still can.

Read the points column as the same weight on a 100-point review, not as a grade. A product that is excellent at UK prep and silent on US entry types can still be the right buy for a British desk. A product that writes beautifully into CargoWise can still be the wrong buy for a Sequoia shop. The disqualifier is a mismatch on the row that carries your actual work, not a low average across rows you do not use.

What each company sells on a customs desk

Wove's forwarder page describes four jobs: quotes from request emails, shipments from pre-alerts, customs entry drafts from commercial documents, and accounts-payable matching against accruals. The customs page says declarations are built for the United States, Canada, the United Kingdom, the European Union, and Australia, with classification rationale kept per line, and with checks for anti-dumping and countervailing duty scope and for other government agency requirements. The same page says a licensed broker reviews and files. That is the right shape for a US broker desk that wants speed without handing the entry summary to an unattended process. Navia Freight is quoted on the forwarder page saying Wove automated quoting and shipment creation on CargoWise. Treat that as the vendor's published customer statement.

Wove's limits sit in what the pages do not sell. It does not present itself as the broker. It does not publish a dollar price. The native story is CargoWise through eAdaptor, including eDocs. Magaya, Logiware, and NetSuite are named, and the page says other systems connect by API. A forwarder on a homegrown operations system should assume an integration project. Tariff depth is a real part of the product, including a public calculator and an API for duty lookup, stacked tariffs, description search, historical rates, and trade-agreement eligibility, described on Wove's January 12, 2026 API announcement. Implementation, on the forwarder FAQ, is a short configuration the vendor handles, with rules set without code. Confirm that timeline against your own branch count before you repeat it internally.

Ripple's customs page describes reading commercial invoices and multi-document packs, classifying against the UK Global Tariff, structuring data for the Customs Declaration Service, and escalating genuine exceptions. The Pentagon Freight Services story is more specific about the handoff. Ripple reads the paperwork, applies commodity, procedure, origin, and license logic, and creates a job in Sequoia. Operators check exceptions and submit to the Customs Declaration Service through the Sequoia and CargoWise path they already know. CargoWise appears there as the downstream record for electronic documents, not as the only system of record. Ripple's home-page description of scope, which this guide will not treat as a price page, also covers quote-to-cash and compliance work. Best fit is a forwarder or broker whose cost sits in document prep and re-keying, especially on UK entries, and who can accept a customized build.

Ripple's limits are the mirror image. The detailed filing story that was opened is a UK story. A US broker who needs entry-type logic, bond checks, and a CargoWise customs declaration as the system of record should make Ripple show that path on the buyer's own files. Public pricing is absent. Trade press describes customized delivery rather than a switch a team flips alone. Ripple's own customs marketing and the Pentagon write-up should be read together: one talks about sending a compliant submission, the other shows a person still checking and sending. Get the submit button in the statement of work.

Our reading of that matrix is narrow on purpose. Wove is the tighter fit when the system of record is CargoWise and the desk wants tariff history in the same vendor. Ripple is the tighter fit when the work is high-variation UK prep plus other operational workflows, and the buyer will pay for a build. Neither cell above is a reason to skip the filer.

Ripple's customs page states a Clearance-cost cut: 60%, a 90 percent drop in rework and variation, clearance in under 3 minutes, and more than 100 complex clearances a day, according to Ripple (2026). Those are the vendor's published claims for its UK customs use case. Ask which customer's file mix produced them before you put them in a business case.

Pentagon's published UK prep numbers

Ripple's Pentagon Freight Services story is the clearest public window into how a forwarder's customs desk actually changes. The prior tool, as Pentagon describes it on that page, needed a separate workflow for each customer. A logic change could touch many workflows, and a bad change could declare the wrong destination country on large entries. The replacement path keeps Sequoia as the entry system and keeps a person on the check.

Ripple's Pentagon Freight Services story reports Weekly prep removed: about 1,400 hours, about 140 declarations a day, prep of 0.5 to 2 hours a declaration, up to about 35 full-time-equivalent of weekly capacity, software cost 40 percent lower than the prior tool, and a prior stack of 50 to 80 customer-specific workflows, according to Ripple (2026). The hours figure is the vendor's case-study arithmetic from that volume and that prep range. It is not a guarantee for a desk with cleaner invoices or a different declaration system.

The useful lesson for a buyer is the control model. Central logic replaced a pile of customer-specific flows. Operators still add missing fields, submit, and handle status messages that come back. If your risk committee needs a human on the submit step, this story supports that design. If a salesperson tells you the person disappears, the story you can read in public does not say that.

Pricing checked against the public pages

Pricing checked October 9, 2026. Wove does not publish a platform price. A request to its pricing address returned a not-found page. Its tariff API post says buyers should ask about volume pricing, and it does not print a rate card. Ripple's public FAQ describes outcome-based pricing aligned to transaction volume and workflow complexity, and it tells the reader to make contact for a figure. No dollar amount appeared on the customs page or the Pentagon story. Both vendors are Quote-based.

A three-year cost cannot be computed from those cells. Anyone who hands you a savings spreadsheet before a written quote is guessing. What you can compare is the shape. Wove's public motion includes a free tariff calculator that trade press describes as a lead path into automation subscriptions, API licensing, and white-label tariff data. Ripple's public motion is a scoped build whose price follows the workflows you turn on. A forwarder that only wants tariff lookup for its own customers has a stranger-accessible Wove tool to try before a contract. A forwarder that wants the prep work taken off a UK desk does not get a comparable self-serve Ripple price.

Put the internal cost next to the quote when it arrives. Count the hours your team still spends on exceptions, the broker review, and the integration owner. Those hours do not vanish because a vendor's case study moved them. They move only if your document mix looks like the mix in the story.

The fee a forwarder still pays CBP

Software choice does not repeal the user fee on a US formal entry. The rows below apply the published formula to illustrative entered values so a desk can see when the floor and the cap bind. The dollar values in the first column are examples for the arithmetic. They are not a shipment from either vendor, and they are not a quote.

Formal entries on that table are a Formal MPF rate: 0.3464% with a floor of $34.58, a cap of $670.86, and a $4.15 surcharge when the entry is filed manually, while informal entries are $2.77, $8.30, or $12.45, and the waterborne rows list harbor maintenance at .125 percent ad valorem, according to CBP (2026). The page was last modified October 9, 2026. On a $5,000 entered value the uncapped product is $17.32, so the floor of $34.58 is the fee. On $100,000 the product is $346.40, inside the band. On $250,000 the product is $866.00, so the cap of $670.86 is the fee. A quoting tool that omits the floor will understate the fee on small formal entries. A tool that omits the cap will overstate it on large ones.

The merchandise value that enters this formula excludes duty, freight, and insurance. A CIF invoice total is the wrong base. Whichever vendor prepares the draft, the broker still has to confirm the entered value before the fee is treated as final.

A published invoice, with the arithmetic

An operations lead can check the arithmetic on Wove's own forwarder-page example before believing a slogan. The page shows 1,984 invoice lines, a grand total of $184,206.31, and a processing time of 26 seconds, and the same page states extraction accuracy of 99.9 percent, more than 35 document types, a usual sync from email to CargoWise in under one minute, a go-live of less than a week, and an ocean rate ingestion cut from 48 hours to 15 minutes for Hellmann, according to Wove (2026). Dividing 1,984 lines by 26 seconds gives about 76 lines a second, and dividing $184,206.31 by 1,984 lines gives about $92.85 a line. That pace is a published illustration of a clean push, not a promise for every skewed phone photo. When extraction finishes, the external API can send the webhook event extraction.completed, and the API document sets a default limit of 60 requests a minute and 10,000 requests a day, according to Wove (2026). A named reviewer still accepts the draft before anyone files. The grand total on that illustration is a commercial-invoice figure under a CIF incoterm, so it is not, by itself, the entered value for the fee table above.

Hold the illustration next to your own worst file. If your 2,000-line invoices arrive as scans with merged cells, ask the vendor to run one of those, not the sample on the website. Time the run from message received to a draft a broker will sign, and count the lines the reviewer sends back. That clock is the only one that belongs in your business case.

Where the draft still needs a person

The gap that remains, after either vendor has read the PDF, is the judgment on the lines that do not agree. A proposed configuration, not a live deployment, can sit on that gap. A pre-alert lands in the shared mailbox with an invoice and a packing list. US Tech Automations can be set to pick up that message, pass the attachments into the document step the forwarder has licensed, and wait for extraction.completed before it opens a review task. Prerequisites are an OAuth client allowed to read documents, an export from CargoWise or the other system of record that the reviewer already trusts, and a named person who accepts or rejects the draft. The output is a packet of extracted lines plus an exception list. A gross-weight break, a missing origin, or a low-confidence code stays in the queue. Nothing is filed from that task. Teams that still design the document path by hand can start from this customs documentation workflow guide and then decide which steps belong in a vendor product.

A second configurable path sits on the carrier bill, where the pain is paying the wrong amount. US Tech Automations can be set so that an invoice file arriving by email is matched to the shipment accrual already stored from the forwarder system, and only a variance outside a tolerance the broker writes down creates a task. Prerequisites are the invoice export, the accrual export, and a finance owner for the tolerance. A person approves any dispute before a payment status changes. This is a design you could configure. It is not a measured customer result, and it is not a claim that a deployment is running today. When the system of record itself is changing, read what KN Swiftlog means for logistics operators before you harden a write-back that assumes last year's screens.

The review task can land on a board the branch already watches. A status moving into a channel follows the same idea as this Monday.com to Slack automation. The prerequisite is still a human owner for the exception, because a message in a channel is not a filing decision.

Stitching the desk in Zapier, Make, or n8n

The real alternative for many desks is not a second specialist product. It is a chain built in Zapier, Make, or n8n, or a small service the internal team maintains. Those tools can support run histories, retries, error branches, and audit evidence when someone configures them. A failed extraction can retry. A branch can park a file when two weights disagree. A log can show who approved a line. That is enough for a low volume of one-document invoices if the team is willing to own the design.

The buyer owns the rest. Observability, idempotency, escalation, access control, and maintenance do not arrive by default. Idempotency matters because a pre-alert that is forwarded twice must not create two shipments. Escalation matters because a silent retry that exhausts itself over a weekend leaves a container with no entry draft. Access control matters because commercial invoices are customer data, and a personal automation account is the wrong place to store them. Maintenance matters because a tariff schedule and a mailbox rule both change, and the person who built the scenario may have left.

A proposed US Tech Automations design can be configured to keep those controls in one place: the mailbox trigger, the wait for a success or failure event, a single shipment key so a replay does not duplicate the job, and a hold that blocks filing until the named reviewer acts. Prerequisites are the vendor API or a stable file export, a credential store the company controls, and a written list of fields that always stop the line. The internal build can do the same work if the company staffs it. The difference is who designs the hold and who gets the call when the hold fails. For a straightforward inbox rule with one spreadsheet log, the internal chain is the simpler tool and should win.

What trade press added on Ripple's other work

Ripple's customs pages are not the whole public record. Trade press describes other workflows the same company is selling into logistics operators, which matters if you are comparing vendors as platforms rather than as a single customs button.

A Houston drayage workflow fell from Manual-time share: 95% to 5% across 10 terminals, Pendragon Freight Services moved 20 people off manual customs work, and JAS Worldwide's deployment spans more than 100 countries and 500 locations, with two-week build cycles, according to FreightWaves (2026). The 95 percent figure is feedback the chief executive relayed from that drayage client about quote-to-cash time, not a customs-entry statistic. The Pendragon figure is the one tied to customs extraction, classification, and submission. The JAS work described in that interview is audit, incident, vendor, and ESG reporting. Use each number for the workflow it belongs to.

Two consequences follow for a forwarder. First, Ripple will try to shape the build to your rules, on a cycle measured in weeks in that interview, which is attractive when your standard operating procedure is unusual and painful when you wanted a fixed product. Second, a group that wants one vendor for customs prep and for safety or vendor compliance can put both conversations in the same scoping call. A group that only wants a CargoWise shipment created from a pre-alert should not pay for that wider brief by accident.

Glossary for the entry desk

• HTS. The US Harmonized Tariff Schedule code on a line. The digits past the international core decide the duty stack, including extra Chapter 99 measures.

• UK Global Tariff. The commodity duty schedule Ripple's customs page says it classifies against for UK entries.

• CDS. The UK Customs Declaration Service, the submission route in Ripple's Pentagon story.

• eAdaptor. The CargoWise connection Wove's forwarder FAQ says it uses to write the shipment, the consol, or the booking and to attach documents.

• Pre-alert. The email, often with a bundled PDF, that arrives before the freight and starts the keying work.

• Entry draft. The prepared declaration a broker or operator reviews. It is not the filed entry.

• MPF. The merchandise processing fee on CBP's user fee table. Confirm the current floor, rate, and cap there before you quote a landed cost.

• CPC. The customs procedure code on a UK entry. In the Pentagon story it is one of the logic fields applied before the operator checks the job.

Mistakes that waste a pilot

Buying the demo file instead of your worst file. A clean invoice with typed lines will flatter any extractor. Bring a scanned packing list, a merged spreadsheet, and a photo of a certificate. Ask what the reviewer sees when two totals disagree.

Treating a case-study percent as your percent. The published cuts belong to the document mix and the wage rate of the desk that reported them. Rebuild the sum with your declarations per day and your prep minutes, then throw out any line you cannot source.

Assuming the write-back is the filing. A shipment created in CargoWise, or a job created in Sequoia, still needs the filer. If the statement of work does not name that person, you have bought a faster way to make a draft and a slower argument about who is responsible.

Ignoring the fee base. A CIF total fed into a US formal-entry fee will misstate the merchandise processing fee. Put "entered value, excluding duty, freight, and insurance" in the test script.

Skipping the replay test. Forward the same pre-alert twice. If two shipments appear, the workflow is not safe for a shared inbox, no matter how fast the first run looked.

Signing a wide compliance build when you needed one customs path. Ripple's public scope is broad. Wove's public scope includes tariff data and payables as well as entries. Turn off the modules you will not staff in the first month.

Frequently asked questions

Does Wove file the customs entry itself?

No. Wove's customs page says a licensed broker reviews and files the draft the software prepares.

The product reads invoices, packing lists, and bills, classifies lines, logs a rationale, and can write the declaration into CargoWise. The filing decision stays with the broker. That is the correct split for a licensed operation, and it should be written into the contract so a later rollout does not blur it.

Does Ripple replace the declaration system?

No. Ripple prepares the entry data, and the Pentagon story shows operators submitting through Sequoia to the Customs Declaration Service.

CargoWise is named as a downstream place for electronic documents in that story. Confirm on your own files whether a US entry would be created inside your broker system or only described beside it. The UK story is the one with public operational detail.

Is there a public price for either product?

No. Both are Quote-based on the pages opened for this article.

Wove's pricing address did not resolve to a rate card, and Ripple's FAQ points buyers to a conversation tied to volume and complexity. Ask for a written quote that names the workflows, the volume band, and what happens if volume doubles. Do not accept a verbal savings percent in place of that page.

Which product is built around CargoWise?

Wove's published path writes shipments, quotes, and entry drafts into CargoWise through eAdaptor.

Ripple's published customs story uses Sequoia for the entry and mentions CargoWise for documents after the fact. A CargoWise-only forwarder should start with Wove's pages and make Ripple demonstrate a CargoWise customs write-back if that is a requirement. A Sequoia desk should do the opposite.

Can a small team copy this with Zapier, Make, or n8n?

Yes, if the team designs retries, a replay-safe shipment key, an error branch, and a review queue, and then owns them.

Those products can keep a run history and an audit trail when configured. They do not know your entry rules until you teach them. A specialist product is justified when the document variation or the tariff logic is too heavy for the team that would maintain the chain. It is not justified when the only need is to move one clean attachment into a spreadsheet.

When should the current customs software stay with no new layer?

Keep the current system when extraction is already clean and the only remaining step is a filing click a trained person makes quickly.

Adding a preparation vendor, or a configurable layer above one, has a cost in credentials, exceptions, and training. If that cost is larger than the keying you would remove, the incumbent wins. Measure a week of actual re-key time before you decide the incumbent has failed.

When NOT to use US Tech Automations. Skip that layer when Wove or Ripple already writes the draft into the system your broker trusts and the only missing piece is a review checkbox inside that same system. Skip it when a single Zapier, Make, or n8n scenario plus a shared mailbox covers a low volume of one-document invoices and your team will maintain the scenario. Skip it when you cannot name an API scope or a file export and a person who owns the exception queue, because the configuration would have nothing safe to call and nobody assigned to stop a bad line.

Which desk should pick which product

Choose Wove if you are a forwarder or broker on CargoWise, you want pre-alerts, quotes, entry drafts, and carrier bills handled around that system, and tariff history is part of the same decision. Make the pilot a real 1,000-line invoice from your archive, and require the broker review step to stay visible. Walk away from the purchase if the write-back is a spreadsheet someone still types in, or if you need the vendor to be the filer.

Choose Ripple if your expensive hours are customs data prep across ugly customer formats, especially on UK declarations into Sequoia, or if you also want quote-to-cash and compliance workflows in the same program. Require the statement of work to say who submits to the Customs Declaration Service and how a logic change is controlled centrally. Walk away if you need a self-serve US tariff calculator with a published price and you will not fund a customized build.

Choose neither specialist, and keep a carefully built Zapier, Make, or n8n chain, when the volume is modest, the documents are consistent, and you have a person who will own retries and access. Choose a configurable review layer above the specialist you do buy only when the specialist stops at a draft and you still need a held task, a replay-safe key, and a named approver. That design has prerequisites and a human stop. It is not a shortcut around the license.

The operations lead who started this search can finish it with three artifacts: a sample of the worst live files, a one-page map of who files, and a written quote. Public pages can tell you what each product claims. They cannot tell you the price, and they cannot sign the entry. For the shape of the review hold above whichever vendor you pick, see how US Tech Automations configures this.