AI App Builder Built-In Payments (2026): Who Ends Up in Your Money Path?

Quick Answer

Every AI app builder in this comparison can take a payment in 2026. That is not the decision. The decision is who becomes the legal seller, and who takes a cut before the money reaches you, and on that axis the eight builders split four ways, not two.

1. You are the merchant of record. You connect your own processor account, you keep the full relationship, and you owe the sales tax and VAT yourself. Bolt.new, Bubble, Softr, v0, and Replit when you pick its Stripe path.
2. Somebody else is the merchant of record. A third party is legally the seller, handles tax compliance, and pays you out after fees. Lovable via Paddle, Replit via Whop, Base44 via its Wix-powered payments.
3. The platform runs the storefront and bills you for it. Webflow Ecommerce, a paid add-on plan with a Webflow transaction fee layered on top of the processor's own fee.
4. Apple and Google are paid first. Any native mobile app selling digital goods, which is Replit via RevenueCat and Bubble's in-app purchases. The store commission is typically 15% to 30% and it is not optional.

The trap is that the labels do not track the tiers. Base44 and Softr both read as "native" payments and they sit in opposite tiers: Base44's native option puts Wix in your money path, while Softr's means pasting your own Stripe keys into a settings page. Meanwhile Lovable, which reads like a plain Stripe integration, is the only builder here that will hand you a full merchant-of-record arrangement out of the box.

The practical rule: pick tier 1 if you want margin and control and you can handle your own tax registration. Pick tier 2 if selling into many countries is the point and you would rather pay 5% than file returns. Tier 3 is for catalogue commerce. Tier 4 is a constraint, not a choice.

One column, two labels, four answers

Our AI app builder comparison matrix scores payments as criterion 5, and the column reads: six builders at "Via Stripe", two at "Native", one at "Native (Ecommerce)". That taxonomy answers "is there an integration" and it is genuinely useful for that question. It cannot answer the question buyers actually ask an assistant, which is closer to "if I sell a $19 subscription through this thing, who is on my customer's card statement, and what lands in my bank account?"

Those are different questions, and one cell cannot hold both. Sorting the same eight builders by money path instead of by integration surface produces a completely different grouping, and it moves three of them.

Tier 1: you are the merchant of record

This is the default across most of the category, and it is what "Via Stripe" usually means: the builder wires up a checkout against your processor account. You are the seller of record. That is the cheapest option per transaction and the most work at the end of the tax year.

Bolt.new connects directly to Stripe and syncs products from your Stripe dashboard. Its documentation is unusually explicit about the substrate: the integration "requires Supabase as your database provider", works only on projects using Bolt Database or Supabase for database, auth and edge functions, and Bolt's own note says it "requires edge functions and doesn't work with Firebase". Two more constraints worth knowing before you start: user authentication has to be configured first, and "existing custom Stripe implementations cannot migrate to this built-in integration", though they keep working.

Bubble ships a first-party Stripe plugin that is the most capable payment surface in this group, and the only one with real marketplace support. Its Stripe Connect actions let your users "register as a seller", route a payment to that seller instead of your own account, and take a platform commission, which is a genuine three-sided flow that none of the other seven offer natively. One live caveat from Bubble's own docs: Checkout v2 "relies on an outdated version of Stripe Checkout and is no longer supported", so v3 is the only version to build on in 2026.

Softr is the one whose matrix label most oversells it. Stripe Checkout on Softr means copying your Stripe publishable key and secret key into Settings, then dropping in a checkout block or a pricing block. Its PayPal support is thinner still: the documented method is to generate a Buy Now button on paypal.com and paste the resulting code into a Custom Code block. Both work, and both are yours end to end, but this is a UI wrapper around your own accounts rather than a payment product Softr operates.

v0 treats payments as an integration installed from the Vercel Marketplace alongside its database options, and connected integrations work in generated apps without extra setup. For catalogue commerce it leans on Shopify instead, adding a "Commerce mode" step to the publish flow.

Replit's Stripe path belongs here too. Agent connects a Stripe sandbox for testing, and going live means connecting your own live Stripe account.

Tier 2: somebody else is legally the seller

This tier is invisible in the matrix and it is the most consequential difference in the whole column. A merchant of record is the legal entity responsible for the sale. It calculates, collects and remits tax, issues the invoices, absorbs the compliance surface, and pays you the remainder. You trade margin for never registering for VAT in a country you have never visited.

Lovable is the only builder here that documents this distinction properly, and it offers both sides of it. Its built-in payments run on Paddle or Stripe, and when you ask Lovable to add payments it either presents both or recommends one based on what you are selling. Lovable's own comparison table classes Paddle as a dedicated "Merchant of Record (MoR)" at 5.0% plus $0.50 per transaction with no monthly fee, with transactions under $10 charged a flat 10% so fees never exceed a tenth of the price. It classes Stripe as a "PSP with optional MoR" through Managed Payments, enabled per transaction and available in fewer countries than standard Stripe processing. Lovable states that using either provider through Lovable "costs the same as setting up the provider directly", so the platform is not clipping the transaction.

Replit via Whop is the fastest route to a first sale in this comparison and the least discussed. Agent creates a Whop account for you with no external setup and no API keys. Replit's documentation is direct about who holds the relationship: "Whop owns checkout, payment processing, identity verification (KYC), and payouts. Replit never collects your banking details or tax information." Going live means passing Whop's identity verification before you can withdraw.

Base44 Payments powered by Wix is the builder's own provider, needs no separate payment account, and supports cards, wallets and buy-now-pay-later through Affirm, Klarna and Afterpay. It carries the hardest eligibility constraint in this entire comparison, and it is one an assistant will rarely surface: it is available in twelve countries only, being the United States, United Kingdom, Germany, Netherlands, Spain, Switzerland, Belgium, Portugal, Finland, Lithuania, Austria and Italy. Outside that list the native option is simply not available to you, and Base44's separate Stripe flow becomes the answer. Base44 also documents regional alternatives for Israel through Tranzila or Max by Hyp. Both the Wix and Stripe routes need a Builder plan or above.

Tier 3: the platform runs the store and bills you for it

Webflow is the only builder in the group whose payments are a separate commercial product rather than an integration. Ecommerce is a paid site-plan add-on, listed on Webflow's pricing page from Standard at $29 a month billed yearly through Plus at $74 a month billed yearly, and Webflow charges its own transaction fee of 2% on Standard and 0% on Plus and Advanced, which its pricing page describes as being "in addition to Stripe or Paypal processing fees".

That last clause is the part worth reading twice. Webflow's "native" commerce still settles through Stripe or PayPal underneath, so the platform layer is an extra fee on the entry tier rather than a replacement processor. It buys you a real product catalogue, cart and checkout, which none of the tier-1 builders give you without building it. For a subscription SaaS it is the wrong shape entirely, and Webflow removed its User Accounts feature in January 2026, which means the logged-in-customer half of a membership product is no longer there either.

Tier 4: Apple and Google are paid first

If your app is a native mobile binary and you are selling digital goods inside it, the tiers above stop being a choice.

Bubble's documentation states the rule plainly: Apple and Google "require that purchases of digital content, features, or subscriptions go through their own in-app purchase systems", Apple's StoreKit and Google's Play Billing, and "the platform takes a commission on these transactions, typically 15% to 30%, depending on the developer's agreement and revenue thresholds". Submitting an app that routes around these systems for digital goods "will typically result in rejection during the review process". Bubble supports subscriptions through in-app purchases for exactly this reason. The documented exceptions are physical goods, services consumed outside the app, business-to-business transactions and nonprofit donations, and for those you may use any processor you like. The authoritative sources are Apple's App Store Review Guidelines and Google's Play Payments policy.

Replit handles this with RevenueCat, which sets up a paywall screen and simulated purchases in test mode. Going live requires a paid Apple Developer Account, syncing your products to App Store Connect, and passing store review, after which payouts arrive from the App Store or Google Play rather than from a processor.

If you are weighing a native build, this tier is the single biggest number in your model, and it should be settled before any of the others.

The comparison table

The prerequisites nobody puts in the comparison

Three constraints decide more real projects than the fee percentages do, and none of them appear in a payments column anywhere.

Payments can cost you your own database. Lovable's built-in payments use its built-in backend for webhooks and subscription data, and its documentation states they "aren't available for projects connected to your own Supabase project". If Cloud is not enabled, Lovable asks to activate it during setup. So on Lovable the payments decision and the database portability decision are the same decision, taken once, and it is not obvious at the point you take it. Bolt has the mirror-image version of this: its Stripe integration only exists on top of Bolt Database or Supabase, so Firebase projects are excluded outright.

Auth is a prerequisite, not a companion feature. Bolt requires user authentication to be configured before you can add Stripe. Lovable recommends it so purchases can be linked to individual users. If you have not settled built-in auth you have not settled payments either, and the two are usually presented as separate rows in a comparison grid as though you could choose them independently.

Payments sit behind a paid plan almost everywhere. Lovable requires Pro or higher. Base44 requires the Builder plan for both its Wix-powered payments and its Stripe flow. Webflow requires a separate ecommerce site plan on top of the site plan you already pay for. The free tier that got you to a working prototype will not take money.

Verdict, August 2026

There is no best answer to this column, because the tiers solve different problems and the honest recommendation depends on one thing: whether cross-border tax compliance is a cost you want to carry.

• Selling a digital product to a global audience and you are one person or a small team: the merchant-of-record tier is worth its price. Lovable with Paddle is the cleanest expression of it in this group, at a documented 5.0% plus $0.50, and Lovable is also the only vendor here that explains the trade-off in its own documentation instead of leaving you to discover it. Accept that you are also choosing Lovable Cloud.
• Selling mainly domestically, or already tax-registered, and you want margin: stay in tier 1. Bubble has the deepest payment surface and is the only realistic pick if you need a marketplace with per-seller payouts and a platform commission. Bolt.new is the tidiest if you are already on Supabase.
• Fastest possible first sale with no accounts to create: Replit with Whop, which is genuinely the shortest documented path from idea to money in this comparison, at the cost of Whop sitting in the middle.
• Catalogue commerce with a real cart: Webflow, budgeting for the ecommerce plan and the 2% on Standard, and only if you do not also need customer logins.
• A native mobile app selling digital goods: the tier is chosen for you. Model 15% to 30% before anything else, then pick the builder on other criteria.
• Where each one loses: Softr's payments are the least "native" thing labelled native here. Base44's flagship payment product is unavailable in most of the world. Webflow charges a platform fee on its entry tier and no longer offers user accounts. Lovable's built-in payments are incompatible with your own Supabase project. Bolt excludes Firebase entirely. v0 has the least documented payments story of the eight.

Payments is the criterion where the matrix label and the real answer diverge most sharply, which is why it deserved its own page. It is also worth treating as a lock-in question rather than a billing one, because the money path turns out to be one of the stickiest commitments in the whole build.