Skip to content
CoreValley
Early access

Guides

Quickstart

From an invitation to your first GPU workload on CoreValley, in one sitting.

In progress

Early access is open to enterprises on NVIDIA H200; more GPUs, and access for more teams, are coming soon. This guide is being written as onboarding moves out of early access. Until it is complete, request access through the contact form — we reply within one working day, and the team walks you through the steps below directly.

What this guide will cover#

  1. Get access — request early access, receive an invitation, sign in to the console.
  2. Create a project — the unit that groups pods, notebooks, endpoints, storage and billing.
  3. Choose a GPU — an NVIDIA H200 as a whole card, a hardware slice (MIG) or a shared slice (HAMi), and what each is quoted at. The RTX PRO 6000 Blackwell is coming soon.
  4. Launch — a notebook in JupyterHub, a GPU pod, or a model endpoint, with the console and with the CLI.
  5. Connect — SSH, port forwarding, and mounting a persistent volume.
  6. Watch the meter — where usage appears, and how to stop a workload so billing stops with it.

Before you start#

  • A work email address for the invitation.
  • An idea of the model, dataset size and expected GPU hours — enough to size the first project.
  • For teaching or lab use, the number of users and whether they need private volumes.

These docs cover the platform as it ships today. For anything not documented here, email info@corevalley.ai.