Setup, answers, and recovery

Help without
the terminal.

Start with the visible message in AI Lab. The launcher checks common prerequisites and offers a safe next action without silently changing your computer.

01 / Quick start

Install once. Open from your desktop.

Windows x64
  1. Run the setup file.

    Keep the default location and allow the installer to create the desktop icon.

  2. Open Heidelberg AI Lab.

    The launcher opens its private dashboard in your browser.

  3. Follow the Podman check.

    If needed, install Podman Desktop, then select Repair Podman inside AI Lab.

macOS Apple Silicon
  1. Open the disk image.

    Drag Heidelberg AI Lab to Applications.

  2. Approve the beta once.

    The build is not notarized, so macOS requires an explicit security approval.

  3. Follow the Podman check.

    If needed, install Podman Desktop, then select Repair Podman inside AI Lab.

Linux
  1. Install Node.js and Podman.

    Use your distribution's package manager. AI Lab requires Node.js 20 or newer.

  2. Download the Linux package.

    Save the versioned .tgz file from the download page.

  3. Install and start AI Lab.

    Run npm install --global ./Heidelberg-AI-Lab-0.1.0-beta.3-Linux-any.tgz, then run ai-lab.

AI Lab does not silently install system software.

Node.js is included in both desktop builds. Podman Desktop, WSL2, permissions, models, and external providers remain separate choices.

02 / Requirements

What your computer needs.

OS

A supported system

Windows x64, macOS on Apple Silicon, or Linux with Node.js 20+. Windows ARM and Intel Macs are not beta targets.

Runtime

Podman Desktop

Required for container apps. An existing working Docker installation can be used as a fallback.

Network

Internet for first downloads

Installers, images, and models need an initial download. Existing apps can then work offline when their configured model or service does.

Storage

Space for tools and models

The exact amount depends on the apps and models you choose. AI Lab reports current disk use.

03 / Common problems

Start with what you see.

01 AI Lab says "Container runtime needed"

Install Podman Desktop if it is missing. Open it once, return to AI Lab, and select Repair Podman. AI Lab asks before creating or starting the local Podman machine.

02 The dashboard did not open

Open Heidelberg AI Lab again from the desktop or Applications icon. The dashboard is served only on 127.0.0.1; it is not a public website. Check whether your browser blocked the new tab.

03 An app remains in "Installing"

Initial images and models can be large. Keep AI Lab open and check your connection and available disk space. Refresh the app state after the download completes.

04 A port is already in use

Open Settings in AI Lab and choose an unused port in the 3xxxx range. Stop the affected app before changing its port, restart AI Lab, then install the app again.

05 I want to remove an app

Use the app's management menu. Removal stops and removes its container without deleting the named volume that holds durable data.

06 AI Lab offers an update

The notice opens this website only after you approve it. AI Lab never downloads or installs updates automatically, and launcher updates do not alter your app data.

04 / Privacy and control

Know where every route goes.

Where does the dashboard run?

On your computer, bound to the loopback interface. Lifecycle requests require a temporary launch token.

Where are credentials stored?

Machine-specific provider credentials remain under your local AI Lab data directory and are never returned by the dashboard API.

Can apps reach the Internet?

Apps may use the model or service routes you configure. Local operation does not mean every third-party app is automatically offline.

Can apps control Podman?

No ordinary app receives the Podman or Docker socket. Lifecycle operations stay in the AI Lab launcher.

Still stuck?

Bring the useful facts.

Before asking for help, note your operating system, AI Lab version, the app name, and the exact message visible in the dashboard. Never share API keys or credential files.

Return to downloads