The Fintech tag on apis.io holds 2,078 providers, and the number that frames it is 3.4%. That is the share of the tag in the agent-native band of our agent readiness measure, against 1.4% for the rest of the catalog. Everything else about this category looks like the market. The mean Kin Score is 22.9 against 21.4, the median is 15.6 against 14.7, and mean agent readiness is actually lower, 11.2 against 11.5. Fintech is not a category that is ahead. It is a category with a deep top and a wide floor, and the two cancel out in the average.
The tag is four things, and the five leaders come from different ones. The helper that reads the catalog splits Fintech into lending and consumer finance at 656 providers, wealth and capital markets at 326, payouts and cross-border at 206, and accounting and spend management at 204, with 686 unassigned. Airwallex leads at 85.2 on the Kin Score, exemplar, with access clarity at 100.0 and developer ergonomics at 96.4 across 14 API pages for cross-border payments, foreign exchange and multi-currency accounts. Plaid scores 82.7 on seven API pages for bank account data and open banking, carried by operational transparency at 94.7. Stripe is a tenth of a point behind at 82.6, with operational transparency and access clarity both at 100.0 across 71 API pages, and discoverability at 53.3 the lowest of the five. Uphold scores 81.4 across 25 API pages for digital assets, cards and payments, with access clarity at 100.0. Qonto, the European business bank, scores 77.0 on 13 API pages covering SEPA transfers, cards and supplier invoices. Airwallex and Stripe also led the payments showcase in September, and that is not a coincidence: Payments is this tag’s closest neighbour by co-occurrence at 0.2767. Contract governance is the soft spot for all five, from 45.5 at Airwallex down to 22.0 at both Stripe and Qonto.
Against the rest of the catalog, the composite says nothing happened. A mean of 22.9 against 21.4 and a median of 15.6 against 14.7 is a category sitting on the market. The band ladder shows how it gets there. Exemplar holds 2.2% of the tag against 1.9% of everyone else, and strong holds 5.5% against 6.0%, so the top is the same size as the market’s. The difference is at the bottom. Minimal holds 31.2% of Fintech against 38.8% of the rest, and emerging holds 33.6% against 23.5%. Fintech companies are less likely to have published almost nothing and more likely to have published a little. On 10,541 APIs against 171,718, that is a category that has been building API surface for fifteen years and still scores like one that has not, because the median member of it has a developer page and not much else.
Agent readiness is where the category splits in two. The mean is 11.2 against 11.5, and the bands pull in opposite directions. Agent-native holds 3.4% of the tag against 1.4%, and human-only holds 60.2% against 54.6%. Both tails are larger than the market’s. The dimensions explain which tail is which. Fintech leads on the operational signals an agent needs to act: idempotency is lit for 12.1% against 4.2%, a described event surface for 18.6% against 7.7%, agent skills for 19.2% against 12.6%, and error semantics for 26.1% against 21.5%. It trails on the surfaces that let an agent find and connect to it: an MCP server for 15.7% against 24.8%, a rate limit signal for 22.2% against 32.5%, a well-known catalog for 2.9% against 8.3%, protected resource metadata for 5.9% against 10.0%, and delegated identity for 9.1% against 12.8%. Dry-run mode is lit for 0.2% against 0.7%, and documented reversibility is 0.0% on both sides.
The five show the gate working. Agent-native requires idempotency and a stable error envelope on top of the score floor, and it reads the provider’s own conventions artifact for coverage. Airwallex at 55.7, Stripe at 49.5 and Plaid at 47.9 are all agent-native. Qonto has the highest agent readiness of the five at 58.5 and is agent-ready, because its conventions artifact declares idempotency coverage as partial, and the gate demotes a key that covers only some writes. Uphold at 46.3 is agent-ready with idempotency dark. The two held at agent-ready are also the only two of the five with an agent card lit, so the pair that has written down who they are for agents is the pair that has not finished writing down how to call them safely.
This is the inverse of the pattern in most categories. The usual shape on apis.io is a tag that has solved connection and not safety: MCP servers everywhere, idempotency nowhere. Fintech has it the other way round. Nearly three times the market’s share of its providers declare an idempotency key, and well under two thirds of the market’s share run an MCP server. Money moved twice is a regulated incident, so the category learned to make a call repeatable before anyone asked it to be reachable by an agent. What it has not done is open the door. 60.2% of Fintech is still human-only, and the 3.4% that is agent-native got there by doing the hard part first.