The short version
These two differ less on capability than on billing philosophy. Both extract line items; AutoEntry charges double credits for it while Dext gates it behind per-supplier toggles and a separate credit pool. AutoEntry's credit model suits lumpy, low volume; Dext's subscription tiers suit steady volume with practice tooling on top. Model your actual document mix before choosing.
| Feature | Dext | AutoEntry |
|---|---|---|
| Pricing model | Subscription plans structured around users and monthly document volume; practice plans priced per client (as of July 2026). source | Credit packs: Bronze 50 credits at $13/month up to Sapphire 2,500 at $469/month, prices excluding VAT (as of July 2026). source |
| Line-item extraction cost | Enabled per supplier via the Extract line items toggle; extracted lines draw on a separate credit pool from the document quota. source | An invoice with line items costs 2 credits instead of 1: line-level capture doubles the per-document price. source |
| Unused capacity | Plan tiers reset with the billing cycle; volume is a plan-size decision. source | Unused credits roll over for 3 months (90 days), then lapse. source |
| Statements and other documents | Processes costs, sales documents, and supplier statements within plan volumes. source | Priced per type: supplier statements 2 credits, bank statements 3 credits per page. source |
| Xero publishing | Publishes to Xero, defaulting to Draft status due to Xero API limits. source | Integrates with Xero, publishing captured and verified documents with attachments. source |
| Tax handling | Imports the tax-rate list from Xero; extracts tax amounts, not rates, with defaults set per supplier. source | Captures tax amounts from documents and maps to the destination platform's tax codes during verification. source |
Where ExpenseFlow fits (our stake, disclosed)
ExpenseFlow competes with both tools. Our position in one sentence: line-item extraction should not be a per-supplier toggle or a plan wall, so we extract and code every line of every document by default, priced per client with a firm-wide pool of document tokens instead of expiring credit packs. The linked comparison pages test that pitch against each tool, source by source.
Two ways of charging for the same job
Dext and AutoEntry both do pre-accounting capture: documents in, verified data out, published to the ledger with the source attached. Capability lists overlap heavily, which is why most firms comparing them end up staring at the pricing pages. The pricing pages reward staring, because the two models fail in different ways.
AutoEntry sells credits. A purchase invoice costs 1 credit, an invoice with line items costs 2, a supplier statement 2, and a bank statement 3 credits per page. Packs run from Bronze, 50 credits for $13 a month, to Sapphire, 2,500 credits for $469, and unused credits roll over for 90 days before lapsing. The model is legible: every document has a price tag. It is also a tax on exactly the documents that carry the most bookkeeping information, since line-level capture doubles the per-document cost, and page-heavy bank statements are the most expensive objects in the system.
Dext sells subscriptions: business plans sized by users and monthly document volume, and practice plans priced per client for firms. Line-item extraction is not a per-document surcharge; instead it is a per-supplier toggle, off by default, drawing on a separate credit pool. The subscription smooths the monthly invoice but moves the friction into configuration: someone has to decide, supplier by supplier, whether lines are worth extracting.
Where the credit model bites, and where it saves
Credits punish two client shapes. The first is the line-item client, the builder or wholesaler whose every invoice doubles in cost. The second is the backlog job: a catch-up project with nine months of statements burns credits at 3 per page, and the pack you bought for it lapses if the project slips past the rollover window.
Credits reward the opposite shape: small, irregular clients. A company with 30 documents a month does not need a plan tier sized for 250, and a firm can hold one AutoEntry pack across a long tail of tiny clients. Dext practice pricing per client covers this too, but the credit maths is easier to justify to a client who barely sends paper.
Configuration debt vs purchasing decisions
The deeper difference is where each product asks for ongoing attention. AutoEntry asks at purchase time: which pack, how many credits, will they expire. Dext asks at setup time and forever after: supplier rules, extraction toggles, publishing defaults. A well-configured Dext practice hums; a half-configured one silently publishes header-level data for suppliers nobody toggled, and the books inherit whatever the defaults guessed.
Both publish to Xero competently, defaulting to draft-style workflows where the platform allows, and both attach source documents for the audit trail. Tax handling is similar in kind: both extract tax amounts from the document rather than deciding treatments, then lean on the ledger’s tax-code list and per-supplier defaults. Neither has a jurisdiction rules engine; the judgment calls stay with the bookkeeper.
Choosing between them
Before committing either way, run one client’s last three months of real documents through the maths. Count invoices, count how many carried lines you actually split in the ledger, count statement pages, and price that mix under both models. Ten minutes with the numbers settles arguments the feature pages cannot, and it usually reveals that your portfolio contains both kinds of client: a steady core that fits subscription tiers and a ragged tail that fits credits.
Pick AutoEntry when volume is lumpy, clients are small, and you want per-document costs you can recite to a client without opening a spreadsheet. Watch the 90-day rollover on anything seasonal.
Pick Dext when volume is steady, line items matter for enough suppliers to justify the toggle work, and the practice wants client-level tooling and rules doing first-pass coding.
If gating line-item capture behind supplier toggles and separate credit pools feels backwards, that instinct is the one ExpenseFlow is built around; see the disclosure box above for our stake and the linked pages for the sourced case.