Skip to content
Hayonix

API reference · Schema

CreateDedicatedEndpointRequest

  • namestringrequired

    Lowercase letters, digits and hyphens. The model ID is `<org slug>/<name>`.

    1–48 characters

  • display_namestring | null

    Name to show people

    1–100 characters

  • inactivity_timeout_secondsinteger

    Stop after this long without a request (0: never; default 3600)

    0–604800

  • startboolean

    Start serving now (default true)

  • deployment_namestring

    Name of the first deployment (default `primary`)

    1–48 characters

  • modelstring | HuggingFaceModelSource | FineTunedModelSourcerequired

    Model ID (`publisher/name`)

  • hardwareEndpointHardwarerequired
    Show 2 fields of hardware
    • skustringrequired

      Hardware type code

      1–64 characters

    • gpu_countinteger

      GPUs per replica on GPU SKUs

      0–8

  • regionstring

    Region code

    1–64 characters

  • engineany JSON
  • engine_optionsEngineOptions
    Show 2 fields of engine_options
    • context_lengthinteger

      Context length in tokens each request may use (default: the model's context length, else the engine's default)

      128–1048576

    • parallelinteger

      Requests one replica runs at once

      1–256

  • disk_gbinteger

    Room for the weights (default 10 on CPU, 200 on GPU)

    1–2000

  • min_replicasinteger

    Default 1

    1–32

  • max_replicasinteger

    Default 1

    1–32

  • Show 1 field of autoscaling
    • target_concurrencyintegerrequired

      Concurrent requests per replica to aim for: replicas = ceil(concurrency / target), between min and max

      1–1024

Example
{
  "name": "support-bot",
  "model": {
    "source": "huggingface",
    "repository": "bartowski/SmolLM2-135M-Instruct-GGUF",
    "file": "Q4_K_M",
    "type": "chat"
  },
  "hardware": {
    "sku": "cpu-2c-4gb"
  },
  "min_replicas": 1,
  "max_replicas": 2,
  "inactivity_timeout_seconds": 3600
}

Used by