Lili and Intuit are the two providers the Accounting tag page ranks first, at 64.5 and 62.7 on the Kin Score, both in the strong band and 1.8 points apart. The tag holds 220 providers and 95 APIs, with a mean of 27.1, so both sit more than 35 points above their own category. The Kin Score measures the public API surface, not the product and not the company, and this pairing is a reminder of why that clause matters: one of these is a bank with an accounting feature set and the other is QuickBooks. The question a reader has is which one to build against, and the answer depends on a layer of the score that most readers have never looked at.
The business. Lili is a U.S. small-business banking platform, with banking services provided by Sunrise Banks, that bundles a checking account, savings, a debit card, invoicing, bill pay, expense categorization, and tax buckets. It reaches the accounting tag through that feature set, and its developer surface at dev.lili.co is about onboarding businesses onto the bank: an Application API that returns a hosted onboarding URL, a Customer Management API, a Webhooks API for onboarding lifecycle events, and a remote OAuth 2.0 protected MCP server at mcp.lili.co exposing 44 read-oriented tools over balances, transactions, invoices, bills, and tax estimates. Intuit’s catalog record is thin, a single sentence, but its API pages are the QuickBooks family, and the recorded features are OAuth 2.0, webhooks, minor versioning, a sandbox environment, multi-currency, and custom fields. Both are hosted, commercial, freemium, self-serve, and neither offers a trial. Intuit documents 5 rate limits and one OAuth scope. Lili’s record carries no rate limit values and no documented scopes, which will matter below.
The APIs. Intuit publishes 11 API pages to Lili’s 2, and 4 OpenAPI files to Lili’s 1, all of them the providers’ own. Intuit’s pages include the QuickBooks Payments API, the QuickBooks Payroll and Time API, the QuickBooks Desktop API, and four entity pages carved from the accounting contract, among them the Intuit Invoices API with 7 operations and the Intuit Payments API with 5. Lili’s Lili API carries 11 operations, and its Lili MCP Server is the larger surface. Two disclosures. Intuit’s 16 Arazzo workflows, 21 examples, and its agentic-access profile of 17 operations were all generated by API Evangelist from the QuickBooks contract, not published by Intuit, so they are not Intuit’s credit and the post does not count them. Lili’s MCP server is first-party, and so is its conformance documentation. On what each company itself ships, Intuit has more contracts and Lili has the one live agent door.
The experience, facet by facet. Each takes three. Lili wins access clarity 84.2 to 50.0, the widest gap in the comparison, and discoverability 81.5 to 50.0, and edges contract governance 18.2 to 13.6. Intuit wins operational transparency 55.3 to 23.7, which is where the undocumented rate limits and the missing status signals cost Lili, contract quality 79.2 to 59.1, and developer ergonomics 82.1 to 66.1. The weights decide it. Contract quality is 0.25 of the composite and developer ergonomics 0.20, against 0.10 for discoverability and 0.12 for governance. Intuit’s three facets carry 0.58 of the weight to Lili’s 0.42, and on the six facets alone Intuit leads, 60.0 to 58.2.
So why does Lili rank first? The conditional layers. Both providers trigger the regulatory facet, but under different regimes: Lili is scored against Banking and Open Finance at 77.2, Intuit against Payments at 54.7. Those layers moved Lili up 6.2 points and Intuit up 2.7, and that 3.5-point swing is more than the 1.8-point lead. Read plainly, Intuit has the better API surface on the six core facets and Lili has the better regulatory posture in its own sector, and the two are not scored over the same regime. That is the asterisk on this ranking, and it is the finding.
Agent readiness is closer than the bands suggest. Lili scores 48.6 and Intuit 36.1, both agent-ready, neither agent-native, because idempotency is unlit on both sides and the agent-native band gates on it. The dimensions split cleanly. Lili lights error semantics, agent skills, protected resource metadata, and dynamic client registration, the OAuth plumbing that lets a stranger’s agent register and connect to mcp.lili.co without a human in the loop. Intuit lights OpenAPI examples and documented reversibility, and its agentic access is our derived artifact. Both document an MCP server, both a rate limit signal, both delegated identity. Neither has an agent card, a dry-run mode, or a well-known catalog. Against the accounting cohort, where 21.2% run an MCP server and 10.6% support dynamic client registration, Lili is unusual on the identity layer and ordinary on safety.
The verdict turns on what you are building. If you are integrating a business’s books, Intuit is the choice and the six-facet base says so: more contracts, better contract quality, better ergonomics, documented rate limits, documented reversibility. If you are building an AI assistant that acts for a small business on its bank account, Lili is the only one of the two with a first-party remote MCP server, and the only one whose identity layer lets that assistant onboard itself. What would move Intuit’s number is the identity layer Lili already has, protected resource metadata and dynamic client registration, plus a first-party agentic-access profile to replace ours. What would move Lili’s is operational transparency, the cheapest facet it is losing: publish the rate limits and the status signals for a surface that already handles the harder parts, and the six-facet base flips without any help from the regulatory layer.