Report Endpoint

The SDK sends rate limit data to this endpoint after each AI API call.

Endpoint#

POST https://delimiter.app/api/report

Authentication#

Project key in the Authorization header:

Authorization: Bearer dlm_your_project_key

Request body#

{
  "app": "my-production-app",
  "provider": "openai",
  "model": "gpt-4o",
  "timestamp": "2025-01-15T14:23:01.456Z",
  "limits": {
    "requests_limit": 10000,
    "requests_remaining": 7342,
    "tokens_limit": 2000000,
    "tokens_remaining": 1456000,
    "reset_requests_ms": 43000,
    "reset_tokens_ms": 12000
  },
  "credits": {
    "credits_limit": 100.00,
    "credits_remaining": 42.50,
    "request_cost": 0.003
  }
}

The credits field is optional — it's included when the provider returns credit/cost headers (e.g., OpenRouter). The limits fields may be null for providers that don't return rate-limit headers, but the report is still sent so the provider is tracked.

## Response

- `200 OK` — report accepted
- `401 Unauthorized` — invalid project key
- `429 Too Many Requests` — report rate limit exceeded

## Behavior

The SDK sends reports asynchronously (fire-and-forget). Reports are not retried on failure. If the endpoint is unreachable, the report is silently dropped. Your application is never affected by reporting failures.