How Routing Works

Every request passes through up to two routing tiers. The gateway tries each tier in order and returns the first successful response.

The two tiers

Request → Tier 1: Operators → Tier 2: Direct Provider → Response

Tier 1: Operator routing

Operators run Blueprints: service templates like the LLM Inference Blueprint. They stake tokens, serve models, and compete on price, latency, and reputation.

• Selected by scoring algorithm: reputation (40%) + latency (30%) + price (30%)
• Discovered automatically from on-chain Blueprint Service Manager contracts
• Can be pinned by blueprint, service, or operator address
• See Operator Routing for the full Blueprint catalog

When it’s used: Default for auto routing mode, required for SpendAuth (on-chain payments).

When it’s skipped: Only in two cases. When a compliance flag is set (ZDR, no-train, or data residency), because operators can’t verify compliance. And when providerOptions.gateway.order is supplied.

Tier 2: Direct provider

The gateway calls the provider API directly using platform credentials (or BYOK credentials).

When it’s used: Default for standard requests, and the fallback when tier 1 fails. Only tier used when compliance routing is active.

Always used for: ZDR requests, no-train requests, BYOK with compliance flags.

Compliance mode

When zeroDataRetention or disallowPromptTraining is set:

Request → Tier 2: Direct Provider (verified only) → Response

Tier 1 is bypassed. The gateway routes only to providers with verified compliance agreements. See Zero Data Retention for the trust model.

Routing control