In 2011, if you wanted to accept payments on your platform, you had two options: build it yourself (months of engineering, PCI compliance, bank relationships, fraud detection) or redirect users to PayPal and lose control of the experience.
Then Stripe launched. Seven lines of code. Payments worked. You never thought about it again.
Stripe didn't make payments easier. It made payments invisible. The complexity — currency conversion, fraud screening, bank settlement, regulatory compliance — was still there. It just wasn't your problem anymore.
Accounting is at the same inflection point today. And nobody has built the Stripe for it yet.
The parallel is exact
Before Stripe, every platform that processed payments built their own payment stack or cobbled together integrations with legacy processors. The result was fragmented, expensive, and unreliable. Payments were a cost centre that distracted from the core product.
Today, every platform that serves businesses faces the same situation with accounting. Your users need their transactions classified. They need their VAT computed. They need their income tax filed. They need financial statements. And right now, the options are:
Build it yourself. Hire tax engineers, regulatory specialists, and local accountants in every country you operate in. Spend two years and millions of euros before you have something production-ready. Then maintain it forever as laws change.
Integrate piecemeal. Connect to Xero's API in the UK, DATEV in Germany, a local provider in France. Build and maintain separate integrations for each country. End up with an inconsistent experience and no unified data model.
Ignore it. Let your users figure it out themselves. Watch them churn to competitors who offer more.
These are the same three options platforms had with payments before Stripe. Build, cobble, or ignore.
What the "Stripe for accounting" actually looks like
The analogy isn't about cloning Stripe's product. It's about applying the same infrastructure model to a different domain.
Single integration. Just as Stripe gives you a single API for payments regardless of payment method, currency, or country, accounting infrastructure gives you a single API for accounting regardless of tax type, jurisdiction, or filing format. You integrate once. The infrastructure handles the rest.
Invisible to the end user. Your users don't see Stripe. They see your checkout flow. Similarly, your users shouldn't see the accounting infrastructure. They should see your platform's tax summary, your platform's filing button, your platform's financial dashboard. Your brand, powered by infrastructure they never interact with directly.
Complexity absorbed by the provider. Stripe handles PCI compliance, bank relationships, currency conversion, and fraud detection. Accounting infrastructure handles jurisdiction-specific tax rules, form generation, regulatory changes, and accountant review. The platform doesn't need to understand the complexity — it just calls the API.
Pricing that scales with usage. Stripe charges per transaction. Accounting infrastructure charges per user or per accounting event. You pay for what you use. No upfront licensing, no per-country fees, no minimum commitments. You grow, the cost grows proportionally.
Where the analogy breaks down (and why it's harder)
Accounting infrastructure is harder to build than payment infrastructure, for reasons that matter:
Payments are real-time. Accounting is periodic. A payment either works or it doesn't, and you know immediately. Accounting produces outputs — tax returns, financial statements — that are evaluated weeks or months later by tax authorities. Errors surface late and compound.
Payments are mechanical. Accounting requires judgment. A payment is a transfer of a defined amount between two defined parties. It's binary. Accounting involves classification decisions that are inherently ambiguous: Is this expense personal or business? Which VAT rate applies? Is this revenue recognisable now or later? AI handles most of this, but a human needs to review the edge cases.
Payments have one regulatory framework per country. Accounting has many. Payments are regulated by central banks and payment regulators. Accounting is regulated by tax authorities, accounting standards boards, professional bodies, anti-money-laundering authorities, and sometimes multiple agencies within a single country. The regulatory surface area is much larger.
Payments don't require a licensed professional. Accounting often does. Anyone can build a payment processor (with the right licences). But in many jurisdictions, filing a tax return on behalf of a client requires being a licensed accountant, tax agent, or enrolled agent. The infrastructure provider needs a network of qualified professionals, not just software.
These differences don't make the infrastructure model wrong. They make it harder. They mean accounting infrastructure needs a component that payment infrastructure doesn't: a network of qualified accountants across every jurisdiction, carrying professional liability, reviewing AI-generated outputs.
Why now
Three things have converged to make this the right moment:
AI is good enough for classification. Five years ago, automated transaction classification was unreliable. Today, large language models can read an invoice, understand the context, and classify it correctly with high accuracy. The classification layer — which was always the bottleneck in automated accounting — now works.
MCP enables dynamic integration. The Model Context Protocol lets AI agents discover and use accounting tools dynamically. A platform's AI assistant can call an accounting MCP server's tools to classify transactions, compute tax, and request accountant review — without the platform engineering team building custom integration logic for each capability.
The self-employed economy has scaled beyond manual accounting. There are an estimated 60 million freelancers in Europe and 70 million in the US. Most of them use platforms — for banking, invoicing, payments, gig work, or contractor management. The platforms they use don't offer accounting. The local accountant model doesn't scale to serve them all. Something has to fill the gap.
Who builds this
The company that builds the Stripe for accounting will need three things that rarely coexist:
Deep accounting expertise. Not just software engineers who read about tax. Actual warranted accountants who understand jurisdiction-specific rules, professional liability, and regulatory requirements. The kind of people who trained at Big Four firms and understand what it means to sign off on a filing.
Modern infrastructure engineering. API-first architecture, deterministic computation engines, MCP server support, multi-tenant data isolation, jurisdiction-specific rule modules that can be updated without redeploying the platform. The kind of engineering that fintech companies do well.
A network of local accountants. In every jurisdiction, someone needs to review and sign off. This can't be faked with AI. It can't be hand-waved with disclaimers. It requires actual relationships with actual accountants who carry actual liability.
The intersection of these three — accounting expertise, infrastructure engineering, and a multi-jurisdiction accountant network — is where the next Stripe-scale infrastructure company will come from.
Accora is building accounting infrastructure for platforms. One API, 30+ jurisdictions, accountant-reviewed. Start at accora.ai
Michael Cutajar, CPA — Founder of Accora.