# Acasia

> Official documentation for Acasia GPU infrastructure. Reserve dedicated bare metal hardware, and operate your clusters, model deployments, and credentials in Acasia Cloud.

- [What is Acasia?](https://docs.acasia.com/what-is-acasia.md): One team that navigates power, location, financing, procurement, installation, and management for a GPU fleet at scale.
- [Bare metal reservations](https://docs.acasia.com/bare-metal/overview.md): Dedicated GPU hardware with root access, held by one organization for a defined term. Reservations are invite-only and arranged through the Acasia sales team.
- [Acasia Cloud](https://docs.acasia.com/cloud/overview.md): The web platform for the capacity your organization holds — telemetry, hardware alerting, credentials, and, where the control plane is installed, clusters and model deployments.
- [Platform architecture](https://docs.acasia.com/platform-architecture.md): How Acasia Cloud is structured — the resource model, cluster lifecycle states, and how desired state flows through the control plane to your GPU hardware.
- [Deploy your first inference endpoint](https://docs.acasia.com/get-started/deploy-inference-endpoint.md): Pick a model, pick a cluster, and have a callable REST endpoint in about five minutes.
- [Create an API key](https://docs.acasia.com/get-started/create-api-key.md): Generate a credential for programmatic access, and store it where it will not leak.
- [Connect to a cluster with SSH](https://docs.acasia.com/get-started/connect-cluster-ssh.md): Generate a key pair, authorize the public half, and open a shell on an active cluster.
- [Connect your IDE to a cluster](https://docs.acasia.com/get-started/connect-ide-to-cluster.md): Use VS Code, Cursor, JetBrains, or Jupyter to edit and run code on remote GPU hardware.
- [Dashboard](https://docs.acasia.com/developer-portal/dashboard.md): The at-a-glance view of your organization's clusters, active model deployments, and resource status.
- [Machines](https://docs.acasia.com/developer-portal/machines.md): Review the GPU hardware allocated to your organization and how much of it is still free.
- [Clusters](https://docs.acasia.com/developer-portal/clusters.md): Deployable allocations of GPU, CPU, RAM, and storage carved from the machines you hold.
- [Inference Models](https://docs.acasia.com/developer-portal/inference-models.md): Put a model from the library onto a cluster and serve it behind a REST endpoint.
- [API Keys](https://docs.acasia.com/developer-portal/api-keys.md): View, create, rotate, and revoke credentials for programmatic access to Acasia services.
- [Settings](https://docs.acasia.com/settings/overview.md): Where to find profile, security, billing, SSH keys, and organization users.
- [Profile](https://docs.acasia.com/settings/profile.md): Change the display name and email address on your account.
- [Security](https://docs.acasia.com/settings/security.md): Change the password on your account.
- [Billing](https://docs.acasia.com/settings/billing.md): Payment methods, credits, and who in your organization is allowed to manage them.
- [SSH Keys](https://docs.acasia.com/settings/ssh-keys.md): Create, review, and delete the public keys authorized to reach your clusters.
- [Users](https://docs.acasia.com/settings/users.md): Review who belongs to your organization and what each person can do.
- [Invite Member](https://docs.acasia.com/settings/invite-member.md): Add a new person to your organization and choose what they are allowed to do.
- [How the Marketplace works](https://docs.acasia.com/marketplace/overview.md): Hardware owners list unused GPU capacity. You choose the amount and the end time, pay, and the control plane is installed at checkout.
- [Rent GPU capacity](https://docs.acasia.com/marketplace/rent-gpu-capacity.md): Choose a listing, set the end time, create your organization, and pay. About five minutes end to end.
- [The control plane](https://docs.acasia.com/marketplace/control-plane.md): Software installed on rented hardware that reports per-GPU state into Acasia Cloud and lets you deploy models and resize clusters from the browser.