Skip to content

LabLink

On-demand GPU teaching labs accessible through the browser.

Run a hands-on software workshop -- students get a full desktop with your software pre-installed, you just share a link.

See It in Action

Students visit a link, enter their email, and get a full cloud desktop with your software ready to go. No installation, no setup -- just a Chrome browser.

How It Works

  1. Deploy an allocator server to AWS (Quickstart)
  2. Spin up VMs from the admin panel before your workshop (Workshop Guide)
  3. Share a link -- students visit the URL, get assigned a VM, and start working

LabLink is software-agnostic. You can adapt it for any software that runs in Docker.

The LabLink CLI is the recommended deployment path: run lablink configure && lablink deploy from your own machine. The template repository is available when you want to manage the OpenTofu and workflows yourself.

Cloud-agnostic via Bring-Your-Own clients

LabLink is not tied to AWS. With the CLI's Bring-Your-Own clients mode, any machine with Docker and an internet connection (to pull LabLink's published packages from PyPI and container images) can serve as a client — lab workstations, on-prem servers, or VMs from any cloud. BYO clients are only available through the CLI; the template-repo path does not support them.

Getting Started

Choose a deployment path, then install its required tools.

View requirements

Deploy LabLink to AWS — pick the template repo path or the CLI path.

Get started

Adapt for your software, run a workshop, configure your deployment.

Adapting for your software

Explore the Docs

  • Adapting for Your Software

    Install your own tutorial software on client VMs and customize the environment.

    Adapting

  • Configuration

    Customize instance types, machine images, DNS, and SSL settings.

    Configuration

  • Workshop Guide

    Step-by-step guide for running a workshop: create VMs, share links, collect data, clean up.

    Workshop Guide

  • Deployment

    Deploy through the template repository's GitHub Actions workflows.

    Template deployment

Resources

  • GitHub - Source code, issues, and contributions
  • Template - Ready-to-use deployment template
  • Support - Report issues or request features

Built by Talmo Lab at the Salk Institute for Biological Studies.