White label AI software, compared for agencies

Every platform in this category advertises white-label. Read five pricing pages and you find the word doing at least five different jobs: removing a badge, unbranding a chat widget, rebranding a mobile app, giving your client a branded dashboard, or covering the whole product.

That gap is where agencies lose money. You buy on the word, discover at the client demo that your logo only reaches one surface, and either absorb the upgrade or explain the other company’s name.

Six questions tell you which one you are actually buying.

What white label AI software means in practice

The intent is simple: software sold to you unbranded, so you put your name on it and sell it to your clients.

The specification is not. “White-label” is a commercial promise, not a technical standard, and no two vendors implement the same surface area. One removes a footer badge. Another gives your client a login page on your domain.

Both are telling the truth. They are describing different products.

The six things that separate a platform from a rebranded widget

These order the rest of the article, and they double as the questions to ask every vendor before you sign.

1. Multi-tenancy — does each client get their own isolated workspace, or are they folders in your account?
2. Billing — can you charge your client directly, or does the vendor bill them?
3. Markup control — do model and integration costs pass through at cost, or with the vendor’s margin already added?
4. Branding depth — how far does your brand actually reach?
5. Vertical templates — is there a starting point per industry, or does every client begin from zero?
6. Integration reach — how many of your client’s existing systems can the agent actually touch?

The first four decide whether you can resell at all. The last two decide how fast and how well.

Multi-tenancy: one workspace per client

The question underneath is what happens at client number twenty.

Native multi-tenancy gives each client an isolated workspace: their own data, their own agents, their own billing, invisible to every other client. Adding one is provisioning, not reconfiguration.

Single-tenant tools give you one account and expect you to organise clients inside it. That works to about five and then stops. Data separation becomes a convention you maintain rather than a boundary the platform enforces, and the first time a client asks whether their data is isolated from your other clients, the honest answer is uncomfortable.

This is the single most common reason an agency outgrows a tool it liked.

Billing: who charges the client

Follow the invoice and you find out who owns the relationship.

You bill the client directly — usually through Stripe Connect or similar — and the platform takes a fee. You set prices, you own the payment relationship, and the spread is yours.

The vendor bills your client and pays you a share. Simpler to start, and it makes you a channel rather than a vendor. Your pricing is bounded by theirs.

You buy at wholesale and resell — the marketplace model. Margin is real but often comes with commitments. Vendasta’s white-label tier, for instance, carries a $499 monthly minimum in wholesale product spend on a one-year contract: a purchasing commitment, not a platform fee.

Markup control on model and integration costs

The cost that grows with success, and the one least visible at signup.

Every AI product has usage costs underneath — model calls, integration calls. Two things decide whether they eat your margin: whether the vendor passes them through at cost or with markup already applied, and whether you can set your own markup on top when you bill the client.

Metering is the same question from the other side. CustomGPT.ai meters queries — 1,000 a month on its entry tier, 5,000 at Premium — so the platform cost rises as your clients use it more. Botpress charges by conversation. Flat platform fees put that variance on the vendor instead of on you.

Neither model is wrong. But a metered platform means your margin narrows exactly when a client is succeeding, which is the worst possible time to renegotiate.

Branding depth: domain, logo, powered-by

Here is where the word collapses. Five platforms, five different answers to “what carries my brand”:

Three details the table cannot hold, and each one changes what you are buying.

CustomGPT.ai is more than a badge removal, and the $499 buys a real chat experience. You replace the default chatbot logo, restyle the widget — bubble colours, backgrounds, patterns — write your own intro text and sample prompt buttons, set the assistant’s tone, style and behavioural limits, and serve the chat from your own URL or a dedicated subdomain. What you do not get is a client-facing dashboard: the branding covers the conversation, not an account your client logs into.

Stammer.ai reaches further than its price suggests. Its Agency tier includes a custom domain, brand customisation, client sub-accounts, a SaaS configurator and billing — which is not “a branded dashboard” but the machinery for running clients as separate accounts and charging them. At $197 that puts it structurally closer to full infrastructure than to a rebranded tool.

Vendasta’s entry tier is not white-label at all. It is labelled “Co-branding with Vendasta”: their brand stays visible to your client.

Voiceflow offers white-labeling in its Agencies tier but does not publish a price for it, which is itself worth knowing when you are budgeting.

White label against private label

Worth separating because the terms get used interchangeably and mean different things.

White label is an existing product, rebranded for you, and typically offered to many resellers at once. Your competitor may be selling the same platform under a different logo.

Private label is built or configured for you specifically and is not resold to others under another name. Rarer, more expensive, usually enterprise.

Almost everything marketed as white-label AI software is genuinely white label. That is not a criticism — shared engineering is why the price works — but it means your differentiation has to come from your niche, your service and your pricing, not from the software being uniquely yours.

Can you white-label ChatGPT?

Not ChatGPT. The consumer product, its interface and its name belong to OpenAI, and no reseller agreement changes that.

What you can do is build on the underlying models through the API and wrap your own product around them. That is what most platforms in this category are doing underneath — the model is a component, and the product is everything built on top: the workspaces, the billing, the integrations, the branding.

Which is the useful reframing. The model is not the moat, and it is not the thing you are reselling. Any vendor claiming to white-label ChatGPT means they white-label their own product built on the API.

Chatbot builders you can rebrand

The entry shelf: build a conversational bot, put your logo on it, embed it on a client site.

Botpress is the reference point. Free to start, strong builder, and a large community. Its white-label tier is $150 per month billed annually and covers the chat widget, not the platform — your client gets an unbranded chat box, not a branded dashboard.

For an agency embedding chat on client websites, that is genuinely enough. For an agency selling a platform, it is a different product. The wall arrives when a client asks to log in and see their own numbers.

Agent platforms with client billing

The shelf where reselling becomes structural rather than cosmetic: isolated workspaces per client, you billing the client directly, and costs you can mark up.

Stammer.ai belongs here and reaches into the next shelf. Its $197 Agency tier includes a custom domain, brand customisation, client sub-accounts, a SaaS configurator and billing — the machinery for running clients as separate accounts rather than folders, and charging them yourself.

Lety.ai sits here too, with branding across the platform from the $97 entry tier. The differences between platforms on this shelf are integration reach and billing mechanics rather than whether branding exists at all.

What defines the shelf is that all three structural questions get a yes. That is a much shorter list than the number of platforms advertising white-label.

Vertical tool suites: SEO, receptionist, marketing

Narrower and often cheaper: one job, done well, rebrandable.

These are worth their place when your agency has a single repeatable offer — reputation for dentists, booking for clinics. You get a shorter path to a working product because the vertical assumptions are already made.

The limits are the same in every case. One job means one revenue line, and expanding means another vendor, another integration and another invoice. And the vertical assumptions that made setup fast make the edge cases hard.

Full agency infrastructure

The test for this shelf is not breadth of features. It is whether your business model can run on the platform — the product you sell, the billing that collects for it, the team seats that deliver it, the client accounts that hold it. If you still need three other systems to operate, it is a tool, not infrastructure.

Vendasta passes that test for agencies selling marketing services to small businesses: CRM, reputation, ads, billing and fulfillment, with AI employees mapped across the customer journey. Resale-ready branding costs more than the entry price suggests — a $499 monthly minimum in wholesale spend on a one-year contract.

Lety.ai passes it for a different business: agencies that sell AI agents and earn on subscription. The product is the agent, the billing runs through Stripe Connect at your markup, clients live in isolated workspaces, and the branding is yours from the entry plan. One platform, one invoice, one thing to operate.

The distinction that matters when choosing between them is not which has more features. It is which one matches what you actually sell. An agency selling reputation management and an agency selling AI agents need different infrastructure, and buying the wrong one means paying for surface you never touch.

What each category costs you in margin

Comparing sticker prices across those four shelves is misleading, because they consume margin in different places.

Rebrandable builders are cheapest up front, and the cost surfaces later as capability: the moment a client wants something the widget cannot do, you are buying a second tool.

Agent platforms cost more at entry and less in aggregate, because the branding and billing you would otherwise assemble are included.

Vertical suites are efficient per job and inefficient across jobs. Three verticals means three subscriptions.

Full infrastructure carries the highest fixed cost and is the only shelf your whole business can run on. Its margin risk is not the fee — it is buying infrastructure shaped for a different business than yours, and paying for surface you never touch.

Then there is the variance. A metered platform moves cost onto you exactly as clients grow; a flat fee keeps it on the vendor. Over a year that difference usually outweighs the entry price you compared in the first place.

How to choose: start from what you bill

Work backwards from the invoice and the shortlist writes itself.

Start from what you bill, not from what you build. Name the offer and its price first. A $500-a-month retainer and a $3,000 implementation want different platforms, and the platform is a cost inside the offer rather than the other way round.

Then check the four structural questions. Isolated workspaces, direct billing, markup control, branding depth. If any is a no, you can rebrand but you cannot really resell — and the difference will surface at client number ten.

Then price the commitment, not the month. Wholesale minimums and annual terms turn a comparable figure into an incomparable one.

Then look at reach. Integrations decide whether the agent resolves the request or only discusses it, and that is what your client renews for.

For agencies that answer those questions and land on wanting the whole stack rather than a tool inside it, that is what Lety.ai is built as. Each client gets an isolated workspace with their own data, agents and billing. Branding runs across the platform from the entry plan — your domain, your logo, no powered-by badges anywhere your client sees. You bill directly through Stripe Connect at your own markup, and agents reach 600+ MCP integrations, which is the layer that decides whether an agent can do the job or only talk about it.

There is no revenue share on your subscriptions. The only platform fee is a transaction fee of 1.9-4.9% on payments processed through Stripe Connect, so the spread between your price and your cost stays yours. The starting points are the white-label platform, the reseller program and pricing.

Pricing and branding terms verified against each vendor’s own pricing page on August 2026. These platforms revise tiers often — confirm current terms before committing.