Powered by NVIDIA DGX Spark
Install on a fresh DGX Spark
Use a fresh DGX Spark with Ubuntu 24.04 ARM64, working Docker Engine + Compose v2, NVIDIA drivers/container support, KVM, Git and Python 3.10+. Allow several hundred GB of temporary free space for models, container builds and the package. Do not run a fresh installation over existing business data.
git clone https://github.com/weerocksllc/teamaibox.git
cd teamaibox
sudo bash install/from-source.shRun these commands on the Spark, not on a desktop laptop. The installer downloads model/build inputs, builds runtime images, creates a complete offline package and installs it. This first source build requires internet access and substantial download/build time; clone alone does not install the product. Docker and the NVIDIA environment are reused, not replaced.
Preflight, download problems and resume
--check performs preflight. If interrupted, keep /var/tmp/teamaibox-source-build and rerun the same command: verified downloads and completed stages are reused. Do not start competing builds or delete the cache. Official sources are the default. Where upstream access is unavailable, use the documented alternate-source options or transfer a verified cache/complete offline package from a connected machine; no mirror guarantees availability. See source installation for logs, --cache-dir, --work-dir and recovery. A changed source tree needs a separate work directory, not a bypassed checkpoint.
sudo bash install/from-source.sh --check
sudo bash install/from-source.shAfter installation
- Open
https://<BOX_LAN_IP>/using the Spark's current LAN address. Verify that its locally generated HTTPS certificate belongs to your device. - Sign in as
admin/admin, change the initial password, then create ordinary accounts. Do not expose the initial credentials to the public internet. - Wait for each AI application to become ready, select a language/theme in the profile menu and begin from the workbench.
- For virtual computers, provide your own authorized Windows/Linux ARM64 ISO and OS license. For an offline target, first prepare the complete package, then follow the package installation guide; source preparation itself is not offline.
Open source, with clear boundaries
Original code and documentation are MIT licensed, not every binary in an installation. Third-party code, model weights, NVIDIA components, firmware, Google services and Windows retain their own terms. The repository does not ship model weights, installed Android/Windows disks, signed-in accounts, ISOs or offline packages. Google compatibility is not Play Protect certification; casting does not cover DRM-protected content or Apple certification. Windows 10 ARM64 is not claimed as supported. Performance and concurrency depend on workloads and hardware.
Build it with us
We welcome bug reports, documentation, translations, accessibility improvements, application integrations and code. Start with an issue describing the user problem, fork the repository, create a focused branch and submit a pull request with relevant checks and screenshots. You do not need to own a Spark to improve docs or isolated components; label hardware-dependent results honestly. Never attach secrets, personal files or signed-in device images. Report vulnerabilities privately through SECURITY.md, not a public issue. No mandatory CLA or paid contributor permission is required.
Contribution workflow ↗Can I run Windows on DGX Spark with Team AI Box?
Team AI Box provides virtual Windows and Linux computers, not a replacement for the Spark host operating system. Use a compatible Windows 11 ARM64 or Linux ARM64 ISO, install through the browser-based virtual display and supply any required operating-system license. Guest application and driver compatibility must be checked separately.
Does cloning the repository install Team AI Box?
No. Run install/from-source.sh on a fresh, supported DGX Spark. The initial online preparation downloads models and dependencies, builds containers and creates the installation package. For an offline target, prepare the complete package on a connected compatible machine first. Keep the build workspace to resume interrupted preparation.