Both are AI-native ledgers where agents draft and a person reviews. What separates them is the books each one was tuned for.
Puzzle and Equated start from the same premise: agents draft the books, a person commits them. Puzzle describes itself as an AI-native ledger doing exactly that. So the usual case for an AI-native ledger — that a system of record is not a system of control — settles nothing between these two.
What separates the two products is narrower and more useful. It is the kind of books each one was tuned for.
Puzzle’s published integrations are the US startup stack: Stripe, Brex, Mercury, Ramp, Deel, Gusto. That list is a description of the product. Transactions arriving from those APIs come in one currency, pre-structured, with the counterparty and the amount already right. In that world the hard part of bookkeeping is categorization and close, and Puzzle is built to do it well.
Our books do not arrive that way. A CAD bank line settles a USD invoice at a rate neither the vendor nor the bank published. A Shopify payout nets sales, refunds, chargebacks and fees into a single number. A GST/HST input tax credit has to survive a CRA question eight months later, which means the receipt has to hang off the posting that claims it, not sit in a folder someone could go find.
None of that is a harder version of Puzzle’s problem. It is a different problem, and it is the one Equated was built for.
Puzzle’s public material does not document multi-currency handling, coverage outside the US, or year-end filing. They do not publish an answer either way, so treat it as an open question and ask them directly if it matters to you. But a ledger’s integration list tells you which books it expects, and Puzzle’s expects US ones.
Reach. Puzzle reports 7,000+ accounting firms and startups on the product, and connects natively to the tools a US startup already runs on. Equated has fewer integrations and far fewer customers, and if you are choosing partly on how many people have already hit the edge cases, that does not favour us.
Firms, too. Puzzle sells to practitioners running many client files at once. That is a real product surface, and we have not built it.
| Capability | Equated | Puzzle |
|---|---|---|
| Double-entry general ledger | Yes | Yes |
| Agents draft, a person reviews | Yes | Yes |
| Documented multi-currency and FX | Yes | Not documented |
| GST/HST and Canadian year-end | Yes | Not documented |
| CPA-reviewed year-end filing | Optional add-on | Not documented |
| Native fintech integrations | Fewer | Stripe, Brex, Mercury, Ramp, Deel, Gusto |
Yes. Puzzle describes itself as an AI-native ledger where agents draft categorization, reconciliation, and close for a person to review, which is the same bet Equated makes. Neither product is an old ledger with automation bolted on. The choice between them is not about the architecture, it is about which books the architecture was tuned for.
Puzzle's public material does not document multi-currency handling, coverage outside the US, or year-end filing, and every integration it publishes is US fintech — Stripe, Brex, Mercury, Ramp, Deel, Gusto. Puzzle does not publish an answer either way, so ask them if it decides your case. Equated is built for USD vendor invoices settling against CAD bank lines, GST/HST input tax credits, and an optional CPA-reviewed T2.
Ask both products what happens to a USD invoice paid from a CAD account, and where the receipt for it ends up. A single-currency US startup on the modern fintech stack will not care about the answer, and Puzzle is the stronger fit. A company billing across the border, or filing in Canada, will find the answer decides the whole thing.
Claims about Puzzle last verified August 20, 2026