Skip to content

Plans & Limits

WhiteBoxXAI offers four ways to run the platform, from a lifetime free account to a fully self-hosted deployment. Your plan determines your monthly API allowance, which premium features you can access, and whether you're on the shared platform or a dedicated workspace.

Plan comparison

Capability Demo Free Business Cloud Enterprise Edition
Price Free $0 forever $599/month Custom (lifetime license)
Best for Exploring the product Evaluation, single projects, hobby use Teams running models in production Regulated organizations & governments
Hosting Shared platform Shared platform Dedicated cloud workspace Your own cloud or data center
Access Read-only Full access Full access Full access
Monthly API calls Unmetered 1,000 Included allowance + pay-as-you-grow Unlimited
Models Preloaded samples Your own Your own Your own
Monitoring, drift & explainability View only
Bias & fairness auditing View only
Governance review boards View only
Risk Register & Trust Score View only
Premium features View only
SSO & RBAC Available
Air-gapped deployment
Support Community Community Dedicated support & SLAs Professional installation & training

What each plan includes

Demo

A shared, read-only showcase preloaded with realistic sample data — models, predictions, drift events, bias audits, and reports. It's the fastest way to see what the platform does without connecting any of your own data. Because it's shared and read-only, you can browse everything, including every premium feature, but you can't create, edit, or delete records. Usage isn't metered.

Free

A lifetime free account with full, hands-on access to monitor your own models, with a monthly allowance of 1,000 API calls. It's the right fit for a single project, a proof of concept, or learning the platform end to end. No credit card required.

The free tier doesn't include premium features, and it has no dedicated workspace or subdomain. If you exceed the monthly allowance, further API requests are paused until the next cycle — your existing data and dashboards stay fully accessible.

Business Cloud

$599/month. Your organization gets a dedicated cloud workspace at its own subdomain (your-company.app.whiteboxxai.com), with its own database, cache, and storage. Isolation is physical, not just logical: your data and your compute are separate from every other customer's.

What's included:

  • Dedicated tenant, data isolation, and your own subdomain
  • More CPU, GPU, and memory for larger models or higher-volume workloads
  • A higher included API allowance, with pay-as-you-grow billing past it
  • All premium features, including AI-driven architecture review
  • GRC and audit logging
  • SSO and RBAC available
  • Priority access to new features and beta programs
  • Dedicated support and SLAs

Signing up

Business Cloud is self-serve. After checkout, your workspace is provisioned automatically — this takes roughly 10 to 30 minutes, since a dedicated environment is being built for you. You'll get a verification email once it's ready, and you'll sign in at your own subdomain rather than the shared platform.

Rolling out

Business Cloud is being rolled out in stages. If the pricing page shows Coming Soon instead of a checkout button, self-serve signup isn't live in your region yet — contact us and we'll get you set up.

How overage billing works

Business Cloud includes a monthly API-call allowance. Unlike the Free plan, going past it doesn't block your requests — they're allowed through and the overage is billed as metered usage on your next invoice. This is deliberate: a production monitoring pipeline shouldn't stop logging predictions because it had a busy month.

Your current usage and remaining included allowance are shown in your account settings.

Enterprise Edition

Built and hosted entirely on your own cloud or data center, for regulated organizations and governments. Priced as a lifetime license with optional yearly maintenance and support.

  • Deployed on your infrastructure — full control over your data security
  • Air-gapped deployment for high-trust environments
  • Maximum CPU, GPU, and memory for your largest models
  • ISO 42001, NIST AI RMF, GDPR, EU AI Act, and CCPA governance features
  • SSO, RBAC, and audit trails
  • All premium features, plus additional features available for regulated industries
  • Professional installation, support, and hands-on training

Contact us to discuss an Enterprise Edition deployment.

Premium features

Three capabilities are gated by plan. Business Cloud and Enterprise Edition include all of them; the Demo lets you view them; the Free plan doesn't include them.

Feature What it does
Architecture review AI-driven review of your model architecture and its governance implications.
Model diff Structured comparison between model versions, including changes tied to pull requests.
Local model inference Live attention-weight extraction from a self-hosted model, rather than the heuristic path available on every tier.

Note

Local model inference is not yet wired to a live endpoint. The tier grants access to the capability; availability will be announced when it ships.

Understanding API usage

Most actions through the SDK or REST API count as API calls — registering models, logging predictions, requesting explanations, and generating exports. A few ways to make the most of your allowance:

  • Sample high-volume models. For a model serving thousands of predictions per day, log a representative sample (1–10%) rather than every prediction. You still get statistically meaningful monitoring. See Getting Started.
  • Batch your logging. Sending predictions in batches costs far fewer calls than one per prediction. The SDK supports local buffering and batch flush.
  • Use offline buffering. The SDK can queue operations locally and sync them later — see Offline Mode.
  • Check your usage anytime. Current usage and remaining allowance are in your account settings.

Need help choosing?

Email support@whiteboxxai.com and we'll help you find the right plan — including higher limits, a dedicated workspace, or a self-hosted deployment.