Key facts
| Branding | Your logo, colors and domain |
| Pricing control | You set retail pricing; Plugsky bills wholesale with volume tiers |
| Customer dashboard | Usage analytics, billing and support tickets in your brand |
| API access | Private OpenAI-compatible endpoint per customer |
| Support model | You own first-line support; Plugsky provides escalation |
| Model access | 30+ models under one API |
| Deployment | Plugsky cloud, VPC, on-prem and air-gapped for regulated customers |
| Product status | Live as an Enterprise engagement |
TL;DR
- Customers see your brand, domain and pricing — not Plugsky's.
- You set retail pricing; Plugsky bills wholesale with volume tiers.
- A branded dashboard handles usage, billing and support tickets.
- Each customer gets a private OpenAI-compatible endpoint and its own keys.
- A startup can launch an AI product on wholesale terms, with branded endpoints and dashboards, without building inference infrastructure or a billing stack.
How it works, step by step
- Define the AI products you will sell and the customers you will sell them to.
- Agree wholesale terms and volume tiers with the Plugsky team.
- Configure branding: domain, logo, colors and email templates.
- Set up tenant isolation with per-customer API keys and limits.
- Wire usage metering into your billing and reporting.
- Onboard a pilot customer and measure usage, latency and support load.
- Scale first-line support with a runbook backed by platform escalation.
Try it yourself
Open the LLM API cost calculator →
Why startups white-label AI
Startups need to ship an AI feature this quarter without hiring an infrastructure team. The best path is usually the managed API first, with a clean migration path to private deployment when a large customer or regulator asks.
A startup can launch an AI product on wholesale terms, with branded endpoints and dashboards, without building inference infrastructure or a billing stack. The customer sees your product; underneath, the platform runs 30+ models, retrieval, agent runs, quotas and reliability engineering that would otherwise be a build programme.
What the white-label program includes
Branding covers your logo, colors and domain. Pricing control means you set retail pricing while Plugsky bills wholesale with volume tiers. A branded customer dashboard presents usage analytics, billing and support tickets. Each customer gets a private OpenAI-compatible endpoint, with tenant isolation and per-customer keys and limits. Plugsky runs the models and infrastructure; you own the customer relationship and first-line support, with platform escalation behind you.
Tenant isolation, deployment and governance
Meter usage per tenant for billing and margin analysis, set per-customer rate limits so one tenant cannot exhaust capacity, and keep keys and data isolated between customers from day one. For regulated customers the same 30+ models can run in your VPC, on-prem or air-gapped instead of a shared plane, so one platform covers several segments. For startups, the platform scales underneath as usage grows while the product, pricing and support promise stay yours.
Launch plan and what you still own
White-label hides the platform, not the responsibility. You remain the contracting party for your customers, so their data protection, service levels and support expectations are yours; the program is an Enterprise engagement, so terms, volume tiers and deployment topology are agreed with the team rather than configured self-serve. A typical launch starts narrow: one industry, one high-value workflow, clear service levels and a support runbook.
Prototype on the free plan with plugsky-micro and plugsky-lite, evaluate larger models with the 14-day full-access trial, and use the pricing page to model retail margin against wholesale tiers.
Honest comparison
| Capability | Plugsky white-label | Reselling a public API | Building your own platform |
|---|---|---|---|
| Branding | Your logo, domain and colors | None | Full control |
| Pricing | You set retail; wholesale tiers | You mark up per token | You set everything |
| Customer dashboard | Branded usage, billing, support | Not included | You build it |
| Model coverage | 30+ models, one API | Vendor's catalogue | You host each model |
| Platform ops | Plugsky runs the platform | Vendor runs it | You run it |
| Time to market | Weeks (Enterprise engagement) | Days but undifferentiated | Quarters |
Frequently asked questions
Can a small team launch a white-label AI product?
Yes, if you have distribution and support capacity: the platform provides models, metering and a branded dashboard, and you set retail pricing against wholesale tiers.
Can customers tell it is Plugsky underneath?
No. The customer-facing experience uses your domain, logo, colors and dashboard, and endpoints are presented under your brand.
How does reseller pricing work?
You set the retail price for your customers and are billed wholesale, with volume tiers as usage grows. Exact rates are agreed with the team; check the pricing page for current retail plans.
Do we get a dashboard for our customers?
Yes. Usage analytics, billing and support tickets are presented in your brand, with tenant separation between customers.
Who handles support?
You provide first-line support and own the customer relationship; Plugsky provides the platform and escalation behind you.
Can it run in our VPC or on-prem?
Yes. The same 30+ models can run on Plugsky cloud, in your VPC, on-prem or air-gapped for customers with strict residency requirements.
Is there a free plan or trial to prototype?
Yes — start on the free plan with plugsky-micro and plugsky-lite, and use the 14-day full-access trial to evaluate larger models before signing wholesale terms.