Ray logo

Ray

Ray is an open-source unified compute framework, stewarded by Anyscale, that scales Python and AI workloads from a laptop to a cluster. It consists of Ray Core (a distributed runtime) and a set of AI libraries (Ray Train, Ray Data, Ray Tune, Ray Serve, RLlib) for training, batch inference, hyperparameter search, and model serving. Ray clusters expose a Dashboard and Jobs REST API on the head node (default port 8265) for submitting jobs, inspecting actors and tasks, and serving deployed applications via Ray Serve HTTP endpoints.

3 APIs 0 Features
Distributed ComputingMachine LearningAI InfrastructurePythonModel ServingOpen SourceCompute

APIs

Ray Jobs REST API

REST API on the Ray head node for submitting, listing, inspecting, and stopping Ray jobs, plus streaming logs. Default base URL is http://:8265/api/jobs/. Open-source...

Ray Dashboard API

Internal REST API powering the Ray Dashboard, exposing endpoints for nodes, actors, tasks, placement groups, runtime environments, and cluster events. Same base URL as the Jobs ...

Ray Serve HTTP API

HTTP interface for invoking models and applications deployed via Ray Serve. Each deployed application is exposed as an HTTP endpoint on the Serve HTTP proxy (default port 8000);...

Resources

🔗
Website
Website
🔗
Documentation
Documentation
👥
GitHub Repository
GitHub Repository
👥
GitHub Organization
GitHub Organization
🔗
Anyscale
Anyscale
🔗
Slack
Slack
🔗
Forum
Forum
📰
Blog
Blog
🔗
Issues
Issues
🔗
License
License

Sources

apis.yml Raw ↑ 
aid: ray
name: Ray
description: >-
  Ray is an open-source unified compute framework, stewarded by Anyscale,
  that scales Python and AI workloads from a laptop to a cluster. It
  consists of Ray Core (a distributed runtime) and a set of AI libraries
  (Ray Train, Ray Data, Ray Tune, Ray Serve, RLlib) for training, batch
  inference, hyperparameter search, and model serving. Ray clusters expose
  a Dashboard and Jobs REST API on the head node (default port 8265) for
  submitting jobs, inspecting actors and tasks, and serving deployed
  applications via Ray Serve HTTP endpoints.
type: Index
image: https://kinlane-productions.s3.amazonaws.com/apis-json/apis-json-logo.jpg
tags:
  - Distributed Computing
  - Machine Learning
  - AI Infrastructure
  - Python
  - Model Serving
  - Open Source
  - Compute
url: >-
  https://raw.githubusercontent.com/api-evangelist/ray/refs/heads/main/apis.yml
created: '2026-05-11'
modified: '2026-05-11'
specificationVersion: '0.19'
apis:
  - aid: ray:jobs-api
    name: Ray Jobs REST API
    description: >-
      REST API on the Ray head node for submitting, listing, inspecting,
      and stopping Ray jobs, plus streaming logs. Default base URL is
      http://<head-node>:8265/api/jobs/. Open-source clusters are typically
      unauthenticated; production deployments rely on network controls or
      Anyscale-managed authentication.
    humanURL: https://docs.ray.io/en/latest/cluster/running-applications/job-submission/rest.html
    baseURL: http://127.0.0.1:8265/api
    tags:
      - Jobs
      - Cluster
      - Submission
      - Logs
    properties:
      - type: Documentation
        url: https://docs.ray.io/en/latest/cluster/running-applications/job-submission/rest.html
      - type: Python SDK
        url: https://docs.ray.io/en/latest/cluster/running-applications/job-submission/sdk.html
      - type: CLI
        url: https://docs.ray.io/en/latest/cluster/running-applications/job-submission/cli.html
  - aid: ray:dashboard-api
    name: Ray Dashboard API
    description: >-
      Internal REST API powering the Ray Dashboard, exposing endpoints for
      nodes, actors, tasks, placement groups, runtime environments, and
      cluster events. Same base URL as the Jobs API (http://<head>:8265).
    humanURL: https://docs.ray.io/en/latest/ray-observability/getting-started.html
    baseURL: http://127.0.0.1:8265/api
    tags:
      - Observability
      - Dashboard
      - Cluster State
      - Actors
    properties:
      - type: Documentation
        url: https://docs.ray.io/en/latest/ray-observability/getting-started.html
  - aid: ray:serve-api
    name: Ray Serve HTTP API
    description: >-
      HTTP interface for invoking models and applications deployed via Ray
      Serve. Each deployed application is exposed as an HTTP endpoint on
      the Serve HTTP proxy (default port 8000); authentication and routing
      are configured per deployment.
    humanURL: https://docs.ray.io/en/latest/serve/index.html
    baseURL: http://127.0.0.1:8000
    tags:
      - Model Serving
      - Inference
      - HTTP
    properties:
      - type: Documentation
        url: https://docs.ray.io/en/latest/serve/index.html
      - type: Production Guide
        url: https://docs.ray.io/en/latest/serve/production-guide/index.html

common:
  - type: Website
    url: https://www.ray.io
  - type: Documentation
    url: https://docs.ray.io
  - type: GitHub Repository
    url: https://github.com/ray-project/ray
  - type: GitHub Organization
    url: https://github.com/ray-project
  - type: Anyscale
    url: https://www.anyscale.com
  - type: Slack
    url: https://www.ray.io/community
  - type: Forum
    url: https://discuss.ray.io
  - type: Blog
    url: https://www.anyscale.com/blog
  - type: Issues
    url: https://github.com/ray-project/ray/issues
  - type: License
    url: https://github.com/ray-project/ray/blob/master/LICENSE
maintainers:
  - FN: Kin Lane
    email: [email protected]