Cerebrium

Cerebrium is a serverless GPU infrastructure platform for real-time AI and ML workloads. Developers package code with the Cortex framework and Cerebrium CLI, then deploy each function as an authenticated REST endpoint on autoscaling GPU/CPU compute billed per second, with streaming, WebSocket, async, and OpenAI-compatible invocation patterns.

5 APIs 0 Features

AIGPUServerlessInferenceML Infrastructure

APIs

Cerebrium Inference / Run Endpoints API

Calls each deployed function as an authenticated POST endpoint at /v4/{project}/{app}/{function}, billed per second of GPU/CPU compute. Supports synchronous JSON, Server-Sent Ev...

Cerebrium Streaming Endpoints API

Streams live model output over a Server-Sent Events (text/event-stream) response from a Python generator that yields data, invoked on the same /run endpoint with an Accept of te...

Cerebrium Async Requests API

Submits long-running inference asynchronously with the async=true query parameter, returning 202 Accepted with a run_id; results are forwarded to a configured webhookEndpoint ra...

Cerebrium App Deployment / Management API

Packages a Cortex project and deploys each function as a persistent REST endpoint via the Cerebrium CLI (init, login, run, deploy), with apps, deployments, scaling, and configur...

Cerebrium Logs / Status API

Surfaces app logs, metrics, and platform status through the CLI (cerebrium logs, cerebrium status), the app dashboard, and the public status page.

Collections

Cerebrium Cortex Inference API

OPEN

Pricing Plans

Cerebrium Plans Pricing

4 plans

PLANS

Rate Limits

Cerebrium Rate Limits

4 limits

RATE LIMITS

FinOps

Cerebrium Finops

FINOPS

Resources

👥

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Cerebrium Cortex Inference API
  version: v4
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Inference
    type: folder
  items:
  - info:
      name: Invoke a deployed function
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/{functionName}
      body:
        type: json
        data: '{}'
    docs: Calls a deployed Cortex function. The JSON body maps to the function parameters. Returns run_id, run_time_ms, and
      result.
  - info:
      name: Invoke run function
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/run
      body:
        type: json
        data: '{}'
    docs: Calls the run function on a deployed app.
- info:
    name: Streaming
    type: folder
  items:
  - info:
      name: Stream function output (SSE)
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/run
      headers:
      - key: Accept
        value: text/event-stream
      body:
        type: json
        data: '{}'
    docs: Streams live model output over a Server-Sent Events (text/event-stream) response from a generator function.
- info:
    name: Async
    type: folder
  items:
  - info:
      name: Submit async run
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/run?async=true
      body:
        type: json
        data: '{}'
    docs: Submits the function for asynchronous execution. Returns 202 Accepted with a run_id; results are forwarded to a
      configured webhookEndpoint.
- info:
    name: OpenAI Compatible
    type: folder
  items:
  - info:
      name: Chat completions
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/{functionName}/chat/completions
      body:
        type: json
        data: '{}'
    docs: OpenAI-compatible chat completions served by a deployed function.
  - info:
      name: Embeddings
      type: http
    http:
      method: POST
      url: https://api.aws.us-east-1.cerebrium.ai/v4/{projectId}/{appName}/{functionName}/embedding
      body:
        type: json
        data: '{}'
    docs: OpenAI-compatible embeddings served by a deployed function.
bundled: true