Transparent pricing

Pay per token.
Nothing else.

No annual contract. No hidden multipliers. Add credits and route immediately. Provider costs are passed through at cost.

Pay-per-token

$0.50/ 1M tokens

+ provider costs passed through at cost

What's included with every router

Each request auto-picks from 35+ models based on complexity, cost, and latency targets.

Automatic retries and provider failover. Your traffic stays up even when providers go down.

Track latency and model distribution across all your routers.

Model selection happens in under 50ms. Total latency is routing + model inference.

Router configs and usage data are fully isolated per account. No cross-tenant access.

Works with any OpenAI-compatible client — LangChain, Vercel AI SDK, or raw fetch.