# Anyscale Docs

> Documentation for Anyscale, the platform for scaling AI and ML workloads with Ray.

Anyscale docs are available as markdown. Append `.md` to any doc URL to get the markdown version.

## Get started

> Getting started with Anyscale: setup, configuration, and platform orientation.

Key pages (see section file for all pages):
- [Get started with Anyscale](/get-started.md): Join an Anyscale organization, set up your environment, and start using the platform for distributed AI workloads.
- [What is Anyscale?](/get-started/what-is-anyscale.md): Learn about the core features that make the Anyscale platform.
- [Platform architecture](/get-started/architecture.md): Understand Anyscale's dual-plane architecture with control plane orchestration and customer data plane deployment.

Questions this section answers:
- How do I get started with Anyscale?
- What is Anyscale?
- What is Ray?
- How do I create my first workspace?

All pages: [/get-started/llms.txt](/get-started/llms.txt)

## Developers

> Products and features for developers building AI applications on Anyscale.

Key pages (see section file for all pages):
- [Anyscale for developers](/overview.md): Learn about key tasks, use cases, products, and features for developers on Anyscale.
- [What are Anyscale services?](/services.md): Learn about using Anyscale services to deploy ML and AI model serving endpoints using Ray Serve.
- [What are Anyscale jobs?](/jobs.md): Run production batch workloads such as model training, batch inference, and data processing with Anyscale jobs.
- [Define a Ray cluster](/configuration.md): Learn about configuring compute resources, dependencies, and other environmental settings for deploying Ray clusters on Anyscale.

Questions this section answers:
- How do I run a batch job?
- How do I serve a model?
- How do I manage dependencies?
- How do I configure a Ray cluster?
- How do I monitor my workloads?
- How do I set up CI/CD with Anyscale?

All pages: [/developers/llms.txt](/developers/llms.txt)

## LLMs and agentic AI

> Serving, fine-tuning, batch inference, RAG, and MCP on Anyscale.

Key pages (see section file for all pages):
- [LLMs and agentic AI on Anyscale](/llm.md): Learn about LLM and agentic AI support and features on the Anyscale platform.
- [Serve LLMs with Anyscale services](/llm/serving.md): Learn how to deploy and scale large language models in production using Ray Serve on Anyscale.
- [Post-training for LLMs on Anyscale](/llm/fine-tuning.md): Learn how to adapt foundation models to your specific applications through continued pre-training (CPT), supervised fine-tuning, RLHF, RLVR, and parameter-efficient methods.
- [Run LLM batch inference on Anyscale](/llm/batch-inference.md): Discover the advantages of running Ray Data LLM batch inference workloads on Anyscale.

Questions this section answers:
- How do I deploy an LLM?
- How do I fine-tune an LLM?
- How do I run batch inference with LLMs?
- What is RAG and how do I build it on Anyscale?
- How do I deploy an MCP server?
- How do I choose a GPU for LLM serving?
- How do I benchmark my LLM deployment?

All pages: [/llm/llms.txt](/llm/llms.txt)

## Admins

> Cloud infrastructure, access management, security, and billing for Anyscale admins.

Key pages (see section file for all pages):
- [Anyscale for admins](/administration/overview.md): Learn about admin and organization owner tasks and products on the Anyscale platform.
- [Introduction to Anyscale clouds](/admin/cloud.md): Learn about Anyscale cloud options and how to configure cloud resources for AWS, Google Cloud, and Kubernetes environments.
- [Deploy Anyscale on Kubernetes](/admin/cloud/kubernetes.md): Learn about deploying the Anyscale operator on Kubernetes services such as AKS, GKE, and EKS.
- [User and access management](/administration/organization.md): Manage users, roles, and permissions in your Anyscale organization.

Questions this section answers:
- How do I set up Anyscale on AWS?
- How do I set up Anyscale on Google Cloud?
- How do I set up Anyscale on Azure?
- How do I deploy Anyscale on Kubernetes?
- How do I configure IAM roles for Anyscale?
- How do I manage users and permissions?
- How do I configure SSO?
- How do I set up billing and budgets?
- How do I configure machine pools?
- How do I access secrets from my cloud provider?

All pages: [/admin/llms.txt](/admin/llms.txt)

## Tutorials

> Hands-on tutorials for jobs, services, LLMs, data processing, and training.

Key pages (see section file for all pages):
- [Anyscale tutorials](/tutorials.md): Learn how to use Anyscale by running tutorials that model common application patterns and best practices.
- [Get started with jobs](/tutorials/submit-a-job.md): Run your first job on Anyscale.
- [Get started with services](/tutorials/deploy-a-service.md): Deploy your first service on Anyscale.
- [Deploy Llama 3.1 8b](/tutorials/deploy-an-llm.md): Deploy Llama 3.1 with Ray Serve LLM.

Questions this section answers:
- How do I run my first job on Anyscale?
- How do I deploy my first service?
- How do I deploy an LLM on Anyscale?
- How do I process data at scale?
- How do I train a model on Anyscale?

All pages: [/tutorials/llms.txt](/tutorials/llms.txt)

## Reference

> API reference, CLI and SDK documentation, release notes, and base images.

Key pages (see section file for all pages):
- [Anyscale API reference](/reference.md): Learn about the Anyscale APIs and navigate to reference documentation for the CLI and SDK.
- [CLI configuration](/reference/quickstart-cli.md): Learn how to install and configure the Anyscale CLI.
- [Get started with the Anyscale SDK](/reference/quickstart-sdk.md): Learn how to install, authenticate, and use the Anyscale Python SDK.
- [Anyscale release notes and support lifecycle](/release-notes.md): Learn about new products and features available on Anyscale, as well as the support lifecycle.

Questions this section answers:
- What CLI commands are available?
- How do I install the Anyscale SDK?
- What changed in the latest release?
- What base images are available?
- What are the API endpoints for jobs?
- What are the API endpoints for services?

All pages: [/reference/llms.txt](/reference/llms.txt)