> ## Documentation Index
> Fetch the complete documentation index at: https://docs.ninjachat.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Ninja Router
> Automatically select models and providers for cost, speed and quality.
Route with ninja/auto
Set your budget and routing strategy. NinjaChat selects an eligible model and provider.
## Routing strategies
Prefer lower-cost models that meet the task requirements.Prioritize more capable models for demanding requests.Prefer faster responses based on measured traffic.Use the router's curated model order for each task.
## Send a routed request
Set your model to `ninja/auto`. Add a routing strategy when you want to override the default.
```typescript theme={null}
const response = await client.responses.create({
model: "ninja/auto",
input: "Review this repository migration plan.",
routing: { strategy: "balanced" },
});
console.log(response.routing.resolved_model);
console.log(response.cost_usd);
```
Create `client` with your NinjaChat API key using the [API quickstart](/quickstart). Each response identifies the model and provider used, the cost, and the request ID.
## Control each request
Limit request costs and choose eligible provider policies.Try another eligible model or provider when a request fails.Run the same request across models to compare the results.Inspect cost, latency, selected models and request traces.
You pay the selected model and provider's published rate for actual usage. See [pricing](/pricing) for current rates.