CentML

CentML is an AI inference optimization platform that serves popular open models through OpenAI-compatible serverless endpoints and lets teams stand up dedicated, autoscaling model-serving deployments and compute clusters. The serverless inference API runs at https://api.centml.com/openai/v1 with Bearer API keys, while a separate platform (control-plane) API manages deployments and clusters.

4 APIs 0 Features

AILLMInferenceServerlessGPU

APIs

CentML Serverless Inference (Chat Completions) API

OpenAI-compatible chat completions and text completions across open models (Llama, Qwen, DeepSeek, GPT OSS and others) running on CentML-optimized serverless GPU endpoints, with...

CentML Models API

Lists the models available on CentML's OpenAI-compatible serverless endpoints, returning model IDs that can be passed to the chat completions and completions endpoints.

CentML Deployments / Endpoints API

Platform (control-plane) API for creating and managing dedicated inference, compute, and job deployments - including create, retrieve, update, status, and logs - that expose mod...

CentML Clusters API

Platform (control-plane) API for listing regional clusters and GPU capacity available to place deployments on, used when selecting hardware and region for dedicated endpoints.

Collections

CentML API

OPEN

Pricing Plans

Centml Plans Pricing

3 plans

PLANS

Rate Limits

Centml Rate Limits

4 limits

RATE LIMITS

FinOps

Centml Finops

FINOPS

Event Specifications

CentML Chat Completions Streaming (HTTP + SSE)

AsyncAPI 2.6 description of CentML's **chat completion streaming** surface. CentML does not publish a WebSocket API. The only asynchronous / event-style transport documented at ...

ASYNCAPI

Resources

👥

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: CentML API
  version: '1.0'
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Chat
    type: folder
  items:
  - info:
      name: Creates a model response for the given chat conversation.
      type: http
    http:
      method: POST
      url: https://api.centml.com/openai/v1/chat/completions
      body:
        type: json
        data: '{}'
    docs: OpenAI-compatible chat completions on CentML serverless endpoints. Set stream=true for Server-Sent Events.
- info:
    name: Completions
    type: folder
  items:
  - info:
      name: Creates a completion for the provided prompt.
      type: http
    http:
      method: POST
      url: https://api.centml.com/openai/v1/completions
      body:
        type: json
        data: '{}'
    docs: OpenAI-compatible text completions on CentML serverless endpoints.
- info:
    name: Models
    type: folder
  items:
  - info:
      name: Lists the models available on CentML serverless endpoints.
      type: http
    http:
      method: GET
      url: https://api.centml.com/openai/v1/models
    docs: Returns the model IDs available to the account on the serverless API.
  - info:
      name: Retrieves a model instance.
      type: http
    http:
      method: GET
      url: https://api.centml.com/openai/v1/models/:model
      params:
      - name: model
        value: ''
        type: path
        description: The ID of the model to retrieve.
    docs: Returns metadata for a single model by ID.
- info:
    name: Deployments
    type: folder
  items:
  - info:
      name: Lists all deployments for the account.
      type: http
    http:
      method: GET
      url: https://api.centml.com/deployments
    docs: Platform control-plane endpoint. Lists inference, compute, and job deployments.
  - info:
      name: Creates a dedicated inference deployment.
      type: http
    http:
      method: POST
      url: https://api.centml.com/deployments/inference
      body:
        type: json
        data: '{}'
    docs: Creates a dedicated, autoscaling inference deployment on a selected cluster and hardware instance.
  - info:
      name: Retrieves an inference deployment by ID.
      type: http
    http:
      method: GET
      url: https://api.centml.com/deployments/inference/:deployment_id
      params:
      - name: deployment_id
        value: ''
        type: path
        description: The ID of the deployment to retrieve.
    docs: Retrieves a single inference deployment.
  - info:
      name: Creates a compute deployment.
      type: http
    http:
      method: POST
      url: https://api.centml.com/deployments/compute
      body:
        type: json
        data: '{}'
    docs: Creates a compute deployment.
  - info:
      name: Gets the status of a deployment.
      type: http
    http:
      method: GET
      url: https://api.centml.com/deployments/status/:deployment_id
      params:
      - name: deployment_id
        value: ''
        type: path
    docs: Returns the current status of a deployment.
  - info:
      name: Returns logs for a deployment.
      type: http
    http:
      method: GET
      url: https://api.centml.com/deployments/logs/:deployment_id
      params:
      - name: deployment_id
        value: ''
        type: path
    docs: Returns log output for a deployment.
- info:
    name: Clusters
    type: folder
  items:
  - info:
      name: Lists the regional clusters available to the account.
      type: http
    http:
      method: GET
      url: https://api.centml.com/clusters
    docs: Platform control-plane endpoint. Lists managed regional clusters for placing deployments.
  - info:
      name: Lists available GPU capacity and hardware instances.
      type: http
    http:
      method: GET
      url: https://api.centml.com/capacity
    docs: Returns GPU hardware instance types and available capacity.
bundled: true