Multi-Model Routing

NOME supports the major frontier model providers (including Anthropic, OpenAI, Google, DeepSeek, and MiniMax) and any OpenAI-compatible BYOK endpoint.

BYOK lets custom or self-hosted models participate alongside hosted providers without giving up the run contract, receipts, or cost tracking.

NOME classifies each request by workload type, complexity, and required capabilities. Simple queries route to fast, cost-efficient models. Complex architecture and analysis tasks route to primary reasoning engines.

Routing is deterministic-first with explicit, receipt-backed provider policy. A user-selected route is never silently replaced.

Each run carries explicit effort budgets: max subagents, max tool calls, and max token budget. These prevent over-investment in simple queries.

Multi-model aggregation is available for high-stakes outputs that benefit from comparison.

Connect any OpenAI-compatible API endpoint as a custom model provider. NOME routes to it using the same run contract, receipts, and cost tracking as built-in providers.

Custom endpoints can be targeted by explicit routing policies for specific workload types. If a custom endpoint is selected and unavailable, NOME reports that route failure instead of silently replacing it.

NOME tracks input and output tokens across all connected models simultaneously. Metrics are normalized into clear visual dashboards with per-provider and per-run breakdowns.

The cost audit surface shows real-time spending, historical usage trends, and per-surface token consumption.

When a provider approaches its rate limit or returns errors, NOME Cloud can use a configured redundant provider only when the selected run policy permits overflow.

Failover is visible to the user and produces audit events. The run contract is preserved across provider switches — receipts track which provider handled each segment, and explicit user picks fail closed rather than switching routes.