>

>

Best Payouts API for Vertical SaaS Platforms That Pay Creators or Contractors

Best Payouts API for Vertical SaaS Platforms That Pay Creators or Contractors

Stripe Connect hands your product 77 verification error codes to explain. A vendor of record hands you one status.

Lumanu

11

/

19

/

2026

Last updated Nov 19, 2026.

Your platform pays contractors on behalf of your customers. Five questions decide what you build on top of a payouts API.

How to evaluate a payouts API for your platform: the short answer

Five questions decide the build. Answer them before you compare features, because each answer rules out providers.

Question

What it decides

What you support

Who gets the vendor setup and payable?

Whether your product renders onboarding, explains failures and chases stalled vendors

A remediation surface and a support queue you keep

Who is the vendor's payment counterparty?

Whose name is on the vendor's tax forms and who the vendor contacts

Re-onboarding every vendor if you change it later

Who handles vendor questions and support?

Where "why haven't I been paid" lands

Support headcount that scales with your customers' vendor counts

What is the money flow?

Whether your customers fund payouts directly or you are in the money flow

Workarounds that complicate the integration

How do your customer funds stay separate?

Whether the provider models multi-tenancy or you build it

Account structure you maintain yourself

The first question is the one that teams underestimate. On Stripe Connect your product owns getting each vendor to a payable state. Under a vendor of record API provider, the provider onboards the vendor.

What is a vendor of record?

Two models exist for paying contractors through a platform, and most engineers have only built one of them.

In the familiar model your platform holds the relationship with each payee. Stripe Connect works this way. Every contractor becomes a connected account under your platform, your application collects their verification details, and your product shows them what to fix when something fails.

With a vendor of record the provider becomes the contractor's counterparty. The contractor signs up with the provider, submits tax forms to the provider, and the provider verifies them, screens them, pays them and files their tax forms. Your platform has one vendor relationship instead of thousands.

The implication for your engineering team: there is no onboarding state to manage, because onboarding is not happening inside your product.

Platforms paying creators via Lumanu this way include Sprout Social Influencer Marketing, Aspire, Tano and Companion, Billion Dollar Boy's creator platform.

Under a vendor of record you can also create and approve a payment for a contractor who has not signed up yet. Lumanu credits the funds, contacts the contractor, and releases the money once they finish. Your integration never gates payment creation on a contractor's state, so the branching logic a Connect integration needs here does not exist.

You give up full control over the relationship with the payee. If your product strategy depends on owning that relationship, read the section on when Stripe Connect fits.

Who gets the vendor to a payable state?

This is the question that decides how much you build, and the two models differ in kind here.

On Stripe Connect your product owns vendor completion. Stripe tells you what a connected account still needs through a requirements hash with seven fields: currently_due, eventually_due, past_due, pending_verification, errors, disabled_reason and current_deadline (Stripe, 2026). A parallel future_requirements hash repeats all seven for requirements coming later. Capabilities carry their own three states of active, inactive and pending.

The work starts when a requirement fails.

77 ways a payee can be stuck

Stripe publishes 77 distinct verification and validation error codes for connected account onboarding: 35 beginning invalid_ and 42 beginning verification_ (Stripe, 2026). They range from invalid_tax_id through verification_document_expired to verification_document_corrupt.

Each one is a moment where a contractor is stuck and your product has to tell them what to fix in words they understand. A creator who uploads a greyscale scan of a passport gets verification_document_greyscale. Turning that into "please upload a colour photo of your ID" is your copy, in your interface, in whatever languages you support.

Your team builds that remediation surface and maintains it as Stripe's requirements change. It rarely shows up in an integration estimate, because the API reference makes the requirements model look like field parsing.

Then you chase

Building the surface is the one-time cost. The permanent cost is that a contractor who stops halfway through onboarding is your problem. Your team notices, your product emails them, and your customer asks you why their creator has not been paid.

Under a vendor of record none of that is yours. Lumanu contacts the contractor, answers their questions and gets them to a state where they can withdraw. Your integration reads a status when you want to display one.

The numbers on the Lumanu side describe the same work from the other end. More than 96% of creators and freelancers finish onboarding without asking anyone a question (Lumanu, 2026). Under 4% contact Lumanu for help, and Lumanu answers them within one business day.

The nuance in this comparison

The full 77-code burden lands on Connect Custom, where your application collects the requirements. Choose Express or Standard and Stripe hosts onboarding, collects the information itself, and your integration redirects contractors through an Account Link instead of rendering the remediation surface.

So the decision has three options to consider:

  1. Connect Custom. You keep the payee experience inside your product and you own all 77 codes.

  2. Connect Express or Standard. Stripe hosts onboarding, which cuts the state burden. The contractor leaves your product and completes the flow under Stripe's branding.

  3. Vendor of record. The contractor onboards with the provider, and your product neither renders the flow nor chases the stragglers.

Pick by what you want to own. A platform that treats payouts as a core surface has an argument for Custom. A platform that treats payouts as plumbing has no reason to build any of it.

Who is the contractor's counterparty?

This decision outlasts the other four.

On Connect the contractor becomes a connected account under your platform. Your platform appears in their onboarding, the account exists in your Stripe dashboard, and the reporting relationship follows that structure.

Under a vendor of record the contractor becomes the provider's vendor. Lumanu verifies them, screens them and files their tax forms. Your customer has one vendor record instead of several hundred, and your platform stays outside the tax relationship.

Three implications to understand:

The filing name. Under a vendor of record, the provider files for the contractors it onboards. Your customer's accounts payable team carries one supplier rather than a list that grows with every campaign.

The support route. Contractors' issues get resolved by whoever onboarded them, so choosing the model determines your support model.

The cost of changing it. Moving a contractor from one model to the other means onboarding them again, collecting tax forms again and re-verifying identity. A platform with thousands of payees cannot do that without every one of them noticing. Settle it before launch.

Who answers contractor questions and issues?

Every payouts integration generates a support queue. Someone has to own it.

On Connect the contractor's relationship runs through your platform, so "why haven't I been paid" arrives in your inbox. Your team then works out whether the answer is a missing document, a failed verification or a payout schedule, which means someone on your side learns the Connect requirements model well enough to explain it to a creator.

That queue scales with your customers' contractor counts rather than with your customer count. A platform that signs ten agencies, each paying four hundred creators, inherits four thousand people who might have a question.

Under a vendor of record the provider answers them. The figures below describe contractors onboarding through Lumanu's own flow, which Lumanu built for creators and freelancers:

Measure

Figure

Creators who finish onboarding without asking a question

More than 96%

Creators who contact support during onboarding

Under 4%, answered within one business day

Payouts that generate a question

0.3%

Source: Lumanu platform data, 2026.

Payout questions are the ones your customer success team cannot answer without access to the payment rail, so they escalate instead of resolving. The 0.3% figure sizes that escalation path.

Read these as Lumanu's own numbers rather than as a rate for the category. They measure how often a contractor needs help getting through one particular onboarding flow. Another provider's flow asks for different information in a different order under different branding, so it would produce its own rate, and nobody else publishes one.

The part that carries over is where that rate lands. Under a vendor of record the provider handles whatever share needs help. On Connect that share arrives at your support desk, at whatever rate your contractors turn out to need it, and you learn that number after launch.

A platform that would rather answer its contractors itself can still do that. Nothing in the model prevents it.

Where the money comes from, and how customers stay separate

Flow of funds

Stripe Connect grew out of marketplaces, where the platform processes payments in and pays out of that flow. If your product takes payment from a buyer and splits it with a seller, the model fits.

Many vertical SaaS platforms do not work that way. Your customer runs a campaign budget in their own bank account and wants to pay creators from it, while your software coordinates the work. You never touch the inbound payment, so there is no processed volume to pay out of.

Platforms in that position end up funding Connect balances through transfers, or asking customers to move money into an account structure built for a different purpose. Those workarounds function, and they account for a good share of the complexity in payouts integrations that run long.

A prefunded model matches that shape. Each customer funds their own balance, your platform creates payments against it, and the provider pays the contractor. Match the model to where your customers' money sits.

Multi-tenancy

One platform, many customers, each paying their own contractors from their own funds. Three questions tell you how much of this you maintain:

Can each customer fund separately? If the provider models one balance per customer, your reconciliation follows the provider's structure. If it models one balance for your platform, you track which money belongs to whom.

Can two customers pay the same contractor? A creator working with four brands on your platform should onboard once. Whether that is possible depends on whether the provider's identity sits with the contractor or with the account that invited them.

Scoping and permissions. Your customers will ask for payment history scoped to them. Either the provider gives you that model or you write it.

When Stripe Connect is the right answer

We build Lumanu, so read this section with that in mind. These are the cases where we would tell you to use Connect.

You already process payments on Stripe. Money arrives in your Stripe balance and you want to pay contractors out of it. Connect keeps that in one system with one reconciliation. Adding a separate payouts provider means a funding step and two ledgers to agree. The integration cost of Connect is real, and so is the cost of running two rails.

You want to own the payee experience. Some platforms treat contractor onboarding as part of the product. If creators completing a profile inside your app is a feature you are building rather than a chore you are absorbing, Connect Custom gives you that control. You get all 77 error codes with it, and that is what the control costs.

You need more than payouts. Treasury, card issuing, charging buyers, a full payments stack under one integration. Connect is a payments platform and a vendor of record is not. If your roadmap includes issuing cards to your contractors, Connect wins on range.

You want the direct relationship with the connected account. Under a vendor of record the payee is the provider's vendor. Platforms that want contractors as their own accounts, with their own balances and their own history inside the platform's system of record, are describing Connect.

You are a traditional 2 sided marketplace for creators and influencers. Buyer pays, platform takes a cut, seller gets the rest, all in one flow. Connect was designed for that. A prefunded model adds a step you do not need.

The pattern underneath: Connect fits when payouts are part of your product, and a vendor of record fits when payouts are infrastructure your product depends on. Vertical SaaS is typically the later, which is why we focus there.

What this means for compliance

The compliance question has its own article, so this is the short version.

Paying contractors creates obligations that bind a legal person: sanctions screening before funds move, a signed tax form on file, a withholding determination, and a year-end filing relationship. Your choice of model decides which of those your company carries.

Keep your platform as the payer and those obligations stay with you or your customer. Use a vendor of record and the provider carries them for the contractors it onboards.

Put one platform-specific question to counsel early: does your platform hold funds at any point? Holding payee funds can bring money transmission rules into scope, and that turns on your flow of funds and your contracts rather than on your code. Ask before you build, since the answer shapes the architecture.

For the detail, see our piece on agentic payments compliance, which works through who carries each obligation and why an automated system cannot carry any of them.

Side by side


Stripe Connect

Vendor of record (Lumanu)

Who gets the contractor setup

Your product, on Custom. Stripe's hosted flow on Express and Standard

The provider

Onboarding states your code handles

7 requirement fields, mirrored in future_requirements, plus 77 error codes on Custom

One status, read only when you want to display it

Can you pay before onboarding finishes

No

Yes. Funds are held and released on completion

Contractor's counterparty

Your platform

The provider

Who files the contractor's tax forms

You or your customer

The provider, for contractors it onboards

Who answers the contractor

You

The provider

Funding model

Pay out of processed volume, or fund balances

Each customer prefunds their own balance

Beyond payouts

Charges, treasury, issuing, a full stack

Payouts and the compliance around them

Trolley, Dots and Airwallex sit between these two positions. Each gives you an API and a tax module while leaving the payee relationship and the filing with you, which places them nearer the Connect column on every row that matters here. We compared all five payout APIs in a separate article on agent-run payouts.

Reference architecture


two payout models · 4 steps each, 3 owned by the platform

two payout models · 4 steps each, 3 owned by the platform

The invite is the same in both models. Everything after it differs. On Connect your product collects requirements, explains failures and chases contractors who stall, and only then can you pay. Under a vendor of record the provider does that work, and you can create the payment at any point.

Frequently asked questions

Should I use Stripe Connect or a vendor of record to pay contractors on my platform?

Use Connect when payouts are part of your product, when you already process payments on Stripe, or when you need charges, treasury or issuing alongside payouts. Use a vendor of record when payouts are infrastructure your product depends on, when your customers fund payments from their own accounts, and when you would rather not build and staff contractor onboarding. The deciding question is whether getting each contractor to a payable state should be your product's job.

What is a vendor of record for payouts?

A provider that becomes the contractor's counterparty. The contractor signs up with the provider, submits tax forms to the provider, and the provider verifies, screens, pays and files for them. Your platform holds one vendor relationship instead of one per contractor.

How long does it take to add contractor payouts to a SaaS product?

It depends on which model you choose and how much of the payee experience you build. Lumanu integrations run from three to four weeks at the fast end to a couple of months, varying by organization size and product surface (Lumanu, 2026). A Connect Custom build adds the remediation surface for 77 verification and validation error codes, which is the part teams underestimate.

Who handles contractor support with embedded payouts?

On Connect, your team does, because the contractor's relationship runs through your platform. Under a vendor of record the provider does. Lumanu reports that more than 96% of creators complete onboarding without contacting anyone, that under 4% ask for help and are answered within one business day, and that 0.3% of payouts generate a question (Lumanu, 2026).

Can I pay a contractor who has not finished onboarding?

Under a vendor of record, yes. Lumanu credits the payment, holds the funds, contacts the contractor and releases the money once they complete signup and tax setup. On Connect a transfer requires an account with the relevant capability active, so onboarding has to finish first.

Do I need a money transmitter license to pay contractors through my platform?

Possibly, if your platform holds payee funds rather than passing them through. The answer turns on your flow of funds and your contracts, and exemptions vary by state. Put it to counsel before you design the money movement, since the answer changes the architecture rather than a setting.

How do I keep each customer's payouts separate?

Ask whether the provider models a funding balance per customer, whether one contractor can be paid by several of your customers without onboarding again, and how payment history is scoped. Your code maintains whatever the provider does not model.

Does my platform or my customer file the contractor's 1099?

It depends on who the contractor's counterparty is. Keep the contractor as your connected account and the filing sits with you or your customer. Under a vendor of record the provider files for the contractors it onboards. Confirm your position with tax counsel before a tax year is underway, because the forms for payments already made are already determined.

Sources and methodology

We counted the Stripe figures from Stripe's published documentation on 18 November 2026. The 77 error codes are the distinct values listed for connected account verification: 35 prefixed invalid_ and 42 prefixed verification_. Connect changes, so check the current pages before relying on a count.

Lumanu figures are first-party platform data from 2026 and describe Lumanu's own operations rather than an industry benchmark. No provider in this category publishes comparable support or onboarding figures.

This article describes architectural trade-offs. It is not legal or tax advice.