Usage-based, with a platform-license alternative.
AlphaBitCore is priced two ways. Pick the shape that matches how your firm wants to procure and budget. Specific credit rates, platform-license sizes, and deployment uplifts are provided on a price sheet during evaluation.
Pay per execution, or license the platform.
Both shapes give you the same platform. The difference is how cost tracks to workload.
Credits per governed execution.
Pricing is tied to per-execution cost. You consume credits as governed executions happen — credits bundle the seven cost dimensions the platform already attests to in the EAC: LLM tokens, tool invocations, external API, compute, storage, risk units, compliance overhead.
Best for firms in evaluation, pilots, or workloads with variable volume. You pay for the work the runtime actually governs.
Annual, fixed fee.
Firms with predictable volume or strict procurement preferences can license the platform on an annual basis. Capacity and deployment footprint are sized to the workload. Usage metering remains available for internal FinOps and chargeback.
Best for firms that have already bounded their governed-AI workload and want pricing certainty across a contract year.
The full platform, either way.
- Gateway, Runtime, Trace Recorder, Event Stream, Contract Registry, Revocation Log, Key Registry
- EAC issuance and signing under customer-managed or platform-managed keys
- Deterministic replay tooling and ValidatePoE verification
- The seven cost-dimension ledger (O1 Budget Closure, O2 Cost Monotonicity)
- Standard integrations with enterprise IdP, service mesh, SIEM, and agent frameworks
- Architecture review, technical evaluation, and security briefing during onboarding
Hosted, customer VPC, or on-premise.
Pricing is the same shape across deployment modes; platform footprint and operational responsibility shift. Deployment uplift, if any, is specified on the price sheet.
Hosted
Runs in AlphaBitCore infrastructure. Fastest to stand up; lowest operational overhead for the customer.
Customer VPC
Deployed inside the customer's cloud account. Data and keys stay in the customer boundary; operational responsibility is shared.
On-premise
Deployed on customer-operated hardware in customer facilities. Required for firms with sovereign-data or air-gap constraints.
Honest clarifications.
Pricing is not tiered by feature. Both shapes give you the full platform: six PEM planes, EAC issuance, deterministic replay, the seven cost-dimension ledger, and every invariant and operational constraint the paper names. Tiering would fracture the governance surface; we don’t do it.
Models and external APIs pass through at cost. LLM-provider tokens, third-party data-feed fees, and external API usage are metered to their real underlying cost. AlphaBitCore does not mark up hosted-model inference. The governance overhead (the attestation layer) is the only thing AlphaBitCore charges for.
Evaluation is not metered. The architecture review, security briefing, research briefing, and technical evaluation phases are at no cost to the firm. You pay when governed executions run in your environment.
Want a price sheet against your workload?
We’ll quote a usage-credit rate or a platform-license size against your specific model inventory, expected execution volume, and deployment choice. Typical turnaround is a week from the architecture review.