Infrastructure and Billing

NOME Cloud — subscribe and receive NOME Credits. We handle API negotiations, rate limiting, failover, and billing. One subscription, all models included.

Sovereign BYOK — connect your own API keys from any supported provider. NOME provides the interface, memory, routing, and orchestration. You pay providers directly.

Enterprise VPC — dedicated backend instances via Azure OpenAI or Google Vertex. WorkOS identity, tenant isolation, custom SLAs, and Nome-owned policy control.

NOME supports keys from Anthropic, OpenAI, DeepSeek, Gemini, MiniMax, and Google Vertex. Each key is validated on entry with real-time provider status checks.

The unified model harness includes credit ledger tracking, BYOK passthrough, and 8 provider adapters. Keys are stored with encrypted persistence through the platform's secret records infrastructure.

NOME Credits are the canonical billing unit for cloud infrastructure usage. Credit summaries include total balance, available credits, usage this period, and included monthly allocation.

The credits system tracks offline savings separately — tasks run on-device using local models consume no credits. BYOK mode routes to providers from your machine with no Nome credits consumed.

Different providers use different tokenization strategies and pricing models. NOME normalizes token counts into comparable metrics across all providers, enabling fair cost comparison and budget-aware routing.

The telemetry dashboard shows input and output tokens for each provider simultaneously, with normalized cost-per-token equivalents.

When one provider approaches its rate limit, NOME seamlessly transitions overflow to a redundant provider. This prevents workflow disruption without manual intervention.

Rate-limit enforcement spans multiple layers: platform quota projection, gateway guardrails, and spine access policy hints. The canonical rate-limit ownership model is consistent across all surfaces.

The harness tracks cumulative token cost per run and surfaces it as a first-class metric. Ensemble mode respects cost budgets defined on the run frame.

Administrators can define per-department routing policies that control which teams access premium models vs cost-effective tiers. Identity meets cost governance.