# List stats URL: https://docs.hanzo.ai/docs/openapi/ai/get-ai-router-stats > Returns the router observability aggregate. `GET /v1/ai/router/stats` > [AI →](https://docs.hanzo.ai/docs/openapi/ai) · [All API references →](https://docs.hanzo.ai/docs/openapi) · [Get started →](https://docs.hanzo.ai/docs/guides) | | | | ------------- | ----------------------------------------- | | **Address** | `https://api.hanzo.ai/v1/ai/router/stats` | | **Method** | `GET` | | **Operation** | `get_ai_router_stats` | | **Auth** | `Authorization: Bearer $HANZO_API_KEY` | Returns the router observability aggregate. Two scopes, one route: - ?scope=platform — PUBLIC-safe aggregate over ALL orgs, no authentication. Emits rates, shares, per-task/per-model counts, throughput, and the cost RATIO (saved\_pct) + counterfactual model id, but NEVER absolute $ levels, org identity, raw events, or feature vectors. This is what world.hanzo.ai polls. - default (org scope) — requires a signed-in principal; scoped to the caller's OWN org (a super admin may pass ?org= to target another org or "" for all). Carries the absolute $/MTok indices for the admin savings panel. Window: ?since= (RFC3339) or ?hours= (default 24, capped). Aggregates only. ## Request `GET /v1/ai/router/stats` takes no parameters and no body — the credential is the whole request. ## Response | Status | Body | Meaning | | ------ | ---- | --------------------------------------------------- | | `200` | | Success. | | `401` | — | No credential, or one this service does not accept. | | `403` | — | A valid credential that may not do this. | `200` body — 46 fields. | Field | In | Type | Always | Description | | ---------------------------------- | ---- | ------------------ | ------ | ------------------------------------------------------- | | `data` | body | any | — | The operation's own result. | | `data2` | body | any | — | A second result the operation defines, when it has one. | | `msg` | body | string | yes | Empty on success, the reason on failure. | | `status` | body | string | yes | One of `ok`, `error`. | | `data.by_model` | body | object | — | | | `data.by_model.*` | body | integer | — | | | `data.by_source` | body | object | — | | | `data.by_source.*` | body | integer | — | | | `data.by_task` | body | object | — | | | `data.by_task.*` | body | ai.TaskStats | — | | | `data.by_task.*.events` | body | integer | — | | | `data.by_task.*.models` | body | object | — | | | `data.by_task.*.models.*` | body | integer | — | | | `data.cost` | body | ai.CostStats | — | | | `data.cost.baseline_model` | body | string | — | | | `data.cost.counterfactual_index` | body | number | — | | | `data.cost.cumulative_saved_index` | body | number | — | | | `data.cost.priced_events` | body | integer | — | | | `data.cost.routed_index` | body | number | — | | | `data.cost.saved_pct` | body | number | — | | | `data.org` | body | string | — | | | `data.quality` | body | ai.QualityStats | — | | | `data.quality.avg_confidence` | body | number | — | | | `data.quality.learned_share` | body | number | — | | | `data.quality.reward_rate` | body | number | — | | | `data.quality.rewarded_events` | body | integer | — | | | `data.quality.shadow_agreement` | body | number | — | | | `data.retrain` | body | ai.RetrainMeta | — | | | `data.retrain.events` | body | integer | — | | | `data.retrain.gate_base` | body | number | — | | | `data.retrain.gate_kind` | body | string | — | | | `data.retrain.gate_metric` | body | string | — | | | `data.retrain.gate_passed` | body | boolean | — | | | `data.retrain.gate_value` | body | number | — | | | `data.retrain.note` | body | string | — | | | `data.retrain.published` | body | boolean | — | | | `data.retrain.trained_time` | body | string | — | | | `data.retrain.version` | body | string | — | | | `data.scope` | body | string | — | | | `data.throughput` | body | ai.ThroughputStats | — | | | `data.throughput.per_hour` | body | integer\[] | — | | | `data.throughput.total_window` | body | integer | — | | | `data.window` | body | ai.StatsWindow | — | | | `data.window.events` | body | integer | — | | | `data.window.since` | body | string | — | | | `data.window.until` | body | string | — | | Failure carries the platform error shape — see [Errors](https://docs.hanzo.ai/docs/errors). ## Examples **CLI** ```bash hanzo ai router stats ``` **SDK** **.ts** ```ts import { Configuration, AiApi } from 'hanzoai'; const api = new AiApi(new Configuration({ accessToken: process.env.HANZO_API_KEY })); const { data } = await api.getAiRouterStats(); ``` **.py** ```python from hanzoai.cloud import ApiClient, Configuration from hanzoai.cloud.api import AiApi client = ApiClient(Configuration(access_token=os.environ["HANZO_API_KEY"])) result = AiApi(client).get_ai_router_stats() ``` **.go** ```go cfg := hanzoai.NewConfiguration() cfg.AddDefaultHeader("Authorization", "Bearer "+os.Getenv("HANZO_API_KEY")) client := hanzoai.NewAPIClient(cfg) resp, _, err := client.AiAPI.GetAiRouterStats(context.Background()).Execute() if err != nil { return err } ``` **.rs** ```rust use hanzo_client::apis::{configuration::Configuration, ai_api}; let mut cfg = Configuration::new(); cfg.bearer_access_token = std::env::var("HANZO_API_KEY").ok(); let result = ai_api::get_ai_router_stats(&cfg, Default::default()).await?; ``` **.java** ```java import ai.hanzo.cloud.ApiClient; import ai.hanzo.cloud.api.AiApi; ApiClient client = new ApiClient(); client.setBearerToken(System.getenv("HANZO_API_KEY")); var result = new AiApi(client).getAiRouterStats(); ``` **HTTP** ```bash curl https://api.hanzo.ai/v1/ai/router/stats \ -H "Authorization: Bearer $HANZO_API_KEY" ``` **MCP** MCP declares no tool for **ai** — `tools/list` on `https://api.hanzo.ai/v1/mcp` names the products it does reach. Use HTTP or an SDK. --- [AI API](https://docs.hanzo.ai/docs/openapi/ai) · [All Hanzo APIs](https://docs.hanzo.ai/docs/openapi) · [Interactive reference](https://docs.hanzo.ai/docs/openapi/reference)