API Reference

Base URL: https://api.proximarun.com/v1

All requests require an Authorization: Bearer <api-key> header. Rate limits: 1000 requests/minute on Growth, 100 on Starter.

POST /infer

POST /v1/infer

Submit an inference request to be routed to the optimal node in your pool. Returns the inference response from the selected node along with routing metadata.

Request body

{
  "prompt": [{ "role": "user", "content": "string" }],
  "latencyBudgetMs": 400,
  "costCeilingUsd": 0.002,
  "reliabilityMin": 0.95,
  "allowColdStandby": false,
  "model": "optional-model-name"
}

Response

{
  "id": "resp_01HX...",
  "content": "The response text from the model...",
  "meta": {
    "nodeId": "edge-us-west-01",
    "latencyMs": 187,
    "costUsd": 0.00082,
    "routingDecisionMs": 8,
    "spilled": false
  }
}

GET /nodes

GET /v1/nodes

List all nodes in your pool, their current health status, and rolling latency statistics.

{
  "nodes": [
    {
      "id": "edge-us-west-01",
      "status": "healthy",
      "p95LatencyMs": 142,
      "requestsLast5m": 841,
      "errorRate": 0.0008,
      "lastCheckedAt": "2026-07-04T10:22:05Z"
    }
  ]
}

POST /nodes

POST /v1/nodes

Add a new node to your pool at runtime, without restarting the client or redeploying your config file.

{
  "id": "new-edge-node",
  "endpoint": "https://new-node.internal/v1",
  "backend": "tgi",
  "costPerToken": 0.0000015,
  "latencyClass": "fast",
  "standby": "hot"
}

DELETE /nodes/{id}

DELETE /v1/nodes/{id}

Remove a node from the routing pool immediately. In-flight requests to that node will complete; no new requests will be dispatched to it.

GET /telemetry

GET /v1/telemetry

Query routing telemetry events. Supports filtering by time range, node ID, and routing outcome. Returns up to 1000 events per page.

Query parameters

Parameter	Type	Description
`limit`	integer	Max events to return (default 100, max 1000)
`nodeId`	string	Filter to a specific node
`from`	ISO8601	Start of time range
`to`	ISO8601	End of time range
`outcome`	string	dispatched, spilled, or failed

GET /status

GET /v1/status

Returns the current operational status of the Proximarun routing engine, API gateway, and telemetry pipeline. Suitable for healthcheck integrations.

{
  "status": "operational",
  "components": {
    "routingEngine": "operational",
    "apiGateway": "operational",
    "telemetryPipeline": "operational"
  },
  "checkedAt": "2026-07-04T10:22:11Z"
}