Key facts
| Branded end-user product | Supported — your app owns the UX and branding |
| API layer | OpenAI-compatible; your product calls it server-side |
| Reseller arrangement | Scoped in a partner or enterprise agreement |
| Self-serve white-label toggle | Not available today |
| Billing flow | Plugsky bills you; you bill your customers |
| Deployment options | Cloud, VPC, on-prem, air-gapped |
| Model metadata | Underlying model names can appear in API responses |
| Product status | Live via agreement |
TL;DR
- Building a branded product on the API is supported and common.
- White-labelling the platform console and billing needs a partner agreement.
- Your customers interact with your product, not your provider — that layer is yours.
- Budget for handling model metadata in your UX; underlying names can surface in responses.
- Plan support, SLA and incident communication as part of the arrangement, not after.
How it works, step by step
- Decide the level: branded application only, or white-labelled platform with your own console.
- Map which surfaces your customers see and where provider metadata could leak into them.
- Choose the deployment that fits your customers: shared cloud, VPC, on-prem or air-gapped.
- Scope the commercial model — flat plan for you, subscription or usage pricing for your customers.
- Confirm support boundaries, SLA flow-down and incident communication in the agreement.
- Build a thin gateway in your backend so model and provider changes never reach end users.
Try it yourself
Two different meanings of white-label
Teams usually mean one of two things, and Plugsky answers them differently.
- Branded product: you ship your own application, sell it under your name, and call the API server-side. This is fully supported — your customers see your brand, and Plugsky is infrastructure they never interact with.
- White-labelled platform: your own console, your own billing, Plugsky invisible end to end. This is a partner or enterprise arrangement scoped case by case, not a switch in the dashboard.
Decide which one you actually need before scoping, because the second is a commercial conversation, not a configuration change.
What we do and what we do not do
What we do: give you an OpenAI-compatible API that is easy to hide behind your own backend, support private deployments (VPC, on-prem, air-gapped) when your customers require isolation, and work with partners on commercial terms. What we do not do: provide a self-serve white-label control panel today, let you rebrand the model providers themselves, or guarantee that underlying model identifiers never appear in API responses. Design your UI and logs so metadata stays in your infrastructure, and read the terms for reseller constraints.
Engineering the abstraction layer
The practical work is a thin gateway in your backend: your product calls your own endpoint, and that endpoint holds the Plugsky key and maps your product's names to model names. This buys three things — you can change models without shipping a client update, you can enforce per-customer usage and rate limits, and you can keep provider metadata out of customer-visible surfaces. It also gives you the seam you need if you later move between deployment models or add a second provider for failover.
Honest comparison
| Capability | Plugsky | Typical API provider | Fully managed white-label platform |
|---|---|---|---|
| Branded end-user product | Supported | Supported | Supported |
| Custom domain and console | Partner/enterprise scoped | Varies | Included |
| Reseller billing | You bill customers, Plugsky bills you | Varies | Platform handles it |
| Hide provider metadata | Your gateway controls it | Varies | Handled by platform |
| Private deployment for clients | VPC, on-prem, air-gapped | Rarely | Dedicated tiers |
| Self-serve start | Yes — API key, two free models | Yes | Usually sales-led |
Frequently asked questions
Can I sell access to Plugsky models under my own brand?
Yes. Build your product on the API and sell it under your brand; the API is designed to sit behind your backend so customers never interact with Plugsky directly.
Can I remove Plugsky branding entirely?
For your end users, yes, by keeping all calls server-side. Removing provider branding from the platform console and billing itself is a partner agreement, not a self-serve option.
Will my customers see the underlying model names?
Possibly, if you pass API responses through. Normalise model identifiers in your gateway if you want the underlying catalogue hidden.
Do you offer reseller pricing?
Commercial terms are scoped in a partner or enterprise agreement. See the live pricing page for the contact path, and bring your volume model with you.
Can I host it for a single enterprise client?
Yes — private deployments (VPC, on-prem, air-gapped) are used exactly for that case and are scoped during onboarding.
Is there an SLA I can pass on to customers?
Plugsky publishes an SLA; a partner agreement sets how commitments flow down to your customers. Review it before you commit to your own guarantees.
How do I start?
Prototype on the free plan with two models, validate quality with the 14-day full-access trial, then contact the team about partner terms.