Skip to content
HALOWERK

Denkwerk

rate limit planner

For each limit the request rate it allows is computed (RPM directly, TPM from input plus counted output tokens, ITPM and OTPM separately, RPD and TPD as daily caps), a safety margin is applied and the binding constraint is reported. Concurrency follows Little's law (rate x latency). Provider rules from the official docs (retrieved 2026-11-01): OpenAI counts the larger of max_tokens and the expected output toward the token limit; Anthropic counts only uncached input toward ITPM on most models.

POST https://denk.halowerk.com/v1/rate-limit-planner

Inputs

Fields of the published input schema
FieldTypeRequiredDescription
providerstringNo
limitsobjectYes
workloadobjectYes
safety_marginnumberNo
backoffobjectNo
Show full input schema
{
  "type": "object",
  "additionalProperties": false,
  "required": [
    "limits",
    "workload"
  ],
  "properties": {
    "provider": {
      "type": "string",
      "enum": [
        "openai",
        "anthropic",
        "generic"
      ],
      "default": "generic"
    },
    "limits": {
      "type": "object",
      "additionalProperties": false,
      "properties": {
        "rpm": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "tpm": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "itpm": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "otpm": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "rpd": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "tpd": {
          "type": "number",
          "exclusiveMinimum": 0
        },
        "max_concurrency": {
          "type": "integer",
          "minimum": 1,
          "maximum": 100000
        }
      }
    },
    "workload": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "input_tokens_per_request",
        "output_tokens_per_request",
        "latency_seconds"
      ],
      "properties": {
        "requests": {
          "type": "integer",
          "minimum": 1,
          "maximum": 1000000000,
          "description": "Total jobs to plan for."
        },
        "input_tokens_per_request": {
          "type": "number",
          "minimum": 1
        },
        "output_tokens_per_request": {
          "type": "number",
          "minimum": 0
        },
        "max_tokens": {
          "type": "number",
          "minimum": 1,
          "description": "OpenAI: counted toward the token limit if larger than the output."
        },
        "cached_input_share": {
          "type": "number",
          "minimum": 0,
          "maximum": 1,
          "default": 0,
          "description": "Anthropic: share of input read from cache (not counted toward ITPM)."
        },
        "latency_seconds": {
          "type": "number",
          "exclusiveMinimum": 0,
          "maximum": 3600
        }
      }
    },
    "safety_margin": {
      "type": "number",
      "minimum": 0,
      "maximum": 0.5,
      "default": 0.1
    },
    "backoff": {
      "type": "object",
      "additionalProperties": false,
      "properties": {
        "base_seconds": {
          "type": "number",
          "exclusiveMinimum": 0,
          "maximum": 60,
          "default": 1
        },
        "max_seconds": {
          "type": "number",
          "exclusiveMinimum": 0,
          "maximum": 3600,
          "default": 60
        },
        "max_retries": {
          "type": "integer",
          "minimum": 1,
          "maximum": 20,
          "default": 6
        }
      }
    }
  }
}

Open original schema

Published input example

{
  "provider": "openai",
  "limits": {
    "rpm": 500,
    "tpm": 200000
  },
  "workload": {
    "requests": 1000,
    "input_tokens_per_request": 1500,
    "output_tokens_per_request": 500,
    "latency_seconds": 4
  }
}

Result & documentation

Machine-readable documentation of this hub

Source of this page: the published HALOWERK catalog. Updated: 2026-11-17 UTC.

More endpoints in this hub

All endpoints in this hub โ†’