The thesis: nearly every payments comparison argues about APIs and dashboards, which is the least important axis. The decision is really about who is legally responsible for sales tax and VAT, and how much of your life you want that to consume. Pick on merchant of record status first, fee percentage second, developer experience a distant third.
The one question that decides it
- Frame merchant of record versus direct processor as the actual fork, and explain what each means in plain terms.
- Argue the trade: an MoR takes a higher cut and handles global tax registration, remittance, and invoices. A direct processor is cheaper and hands you a compliance obligation that grows with every country you sell into.
- State the heuristic: solo founders and small teams selling globally should default to an MoR until the fee difference exceeds the cost of a tax practice.
What billing actually involves
The visible part
- Checkout, plans, upgrades, downgrades, cancellation.
The part that breaks in production
- Webhook delivery and replay, subscription state drift between processor and database, proration, failed payments and dunning, refunds, and the tax invoice a customer's finance team demands.
- Argue this second list is where most of the engineering time goes, and where boilerplates earn their price. Reference /what-is-a-saas-boilerplate.
The options, honestly characterized
Stripe
- Strength: best API, widest ecosystem, most boilerplate support, most agent training data.
- Cost: you are the merchant of record, so global tax is yours. Stripe Tax reduces but does not remove the burden.
Lemon Squeezy
- Strength: merchant of record, simple setup, good for digital products and one-time sales.
- Cost: higher effective fees, less flexible for complex or usage-based billing.
Polar
- Strength: merchant of record aimed at developer products, modern API, strong fit for small SaaS and open source monetization.
- Cost: younger ecosystem, thinner boilerplate and agent coverage.
Paddle
- Strength: merchant of record with enterprise and B2B invoicing maturity.
- Cost: heavier onboarding and approval process, less indie-friendly.
Pricing model before processor
- Argue you cannot pick the processor until you know whether you are selling flat subscriptions, seats, usage, or one-time licenses.
- Note that usage-based billing narrows the field sharply and should be declared early.
- Suggest modeling the numbers before committing, and mention the pricing calculator under /free-tools as a starting point.
What generated billing code gets wrong
- Name the specific failure patterns: webhooks treated as reliable and ordered, no idempotency key, subscription state read from the client, and grant-on-checkout instead of grant-on-webhook.
- Argue the reason agents fail here is that the happy path demos perfectly and the failure modes only appear under retries and race conditions.
- Give the rule: entitlement is derived from processor state stored in your database, written by an idempotent webhook handler, and nowhere else.
- Encode that rule in the agent brief. Link /templates/agents-md and /free-tools/agents-md-generator.
Testing the money path properly
- The minimum: a real test-mode purchase, a replayed webhook, a failed payment, a cancellation, and a resubscribe.
- Argue this suite is the highest-value test in a SaaS codebase and the one most likely to be missing.
- Suggest asking your coding agent to explain how a kit keeps subscription state in sync, then verifying the answer against the files. Reference /blog/ai-agent-ready-boilerplate-checklist.
How to pick a boilerplate on payments
- Argue kits differ far more on billing depth than on auth, and list the differentiators: proration, seat-based plans, usage metering, customer portal, invoice access.
- Point at /categories/Agent-Ready for kits scored on structure and /blog/best-nextjs-boilerplate-2026 for stack-specific shortlists.
Frequently Asked Questions
Should I use a merchant of record or Stripe directly?
- Answer with the threshold: MoR until tax handling is cheaper to insource than the fee delta, and describe roughly where that lands.
Can an AI agent build my Stripe integration?
- Answer: it can wire checkout, but the webhook and entitlement layer needs the pattern specified up front and verified by test.
What is the most common billing bug in a new SaaS?
- Answer: access granted at checkout rather than on webhook confirmation, so refunds and failed payments leave paid access in place.
Do I need usage-based billing on day one?
- Answer no in most cases, and explain the cheap way to leave the door open without building it.
How much do payment fees actually matter?
- Answer: less than founders think at small scale, and give the point at which the percentage starts dominating other concerns.
